Adding upstream version 6.6.15.upstream/6.6.15

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

30 files changed, 15347 insertions, 0 deletions

diff --git a/drivers/media/platform/mediatek/vcodec/decoder/Makefile b/drivers/media/platform/mediatek/vcodec/decoder/Makefile
new file mode 100644
index 0000000000..904cd22def
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/Makefile
@@ -0,0 +1,25 @@
+# SPDX-License-Identifier: GPL-2.0
+
+obj-$(CONFIG_VIDEO_MEDIATEK_VCODEC) += mtk-vcodec-dec.o \
+				       mtk-vcodec-dec-hw.o
+
+mtk-vcodec-dec-y := vdec/vdec_h264_if.o \
+		vdec/vdec_vp8_if.o \
+		vdec/vdec_vp8_req_if.o \
+		vdec/vdec_vp9_if.o \
+		vdec/vdec_vp9_req_lat_if.o \
+		vdec/vdec_av1_req_lat_if.o \
+		vdec/vdec_h264_req_if.o \
+		vdec/vdec_h264_req_common.o \
+		vdec/vdec_h264_req_multi_if.o \
+		vdec/vdec_hevc_req_multi_if.o \
+		mtk_vcodec_dec_drv.o \
+		vdec_drv_if.o \
+		vdec_vpu_if.o \
+		vdec_msg_queue.o \
+		mtk_vcodec_dec.o \
+		mtk_vcodec_dec_stateful.o \
+		mtk_vcodec_dec_stateless.o \
+		mtk_vcodec_dec_pm.o \
+
+mtk-vcodec-dec-hw-y := mtk_vcodec_dec_hw.o
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c
new file mode 100644
index 0000000000..91ed576d68
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.c
@@ -0,0 +1,1020 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ *         Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#include <media/v4l2-event.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "mtk_vcodec_dec_drv.h"
+#include "mtk_vcodec_dec.h"
+#include "vdec_drv_if.h"
+#include "mtk_vcodec_dec_pm.h"
+
+#define DFT_CFG_WIDTH	MTK_VDEC_MIN_W
+#define DFT_CFG_HEIGHT	MTK_VDEC_MIN_H
+
+static const struct mtk_video_fmt *
+mtk_vdec_find_format(struct v4l2_format *f,
+		     const struct mtk_vcodec_dec_pdata *dec_pdata)
+{
+	const struct mtk_video_fmt *fmt;
+	unsigned int k;
+
+	for (k = 0; k < *dec_pdata->num_formats; k++) {
+		fmt = &dec_pdata->vdec_formats[k];
+		if (fmt->fourcc == f->fmt.pix_mp.pixelformat)
+			return fmt;
+	}
+
+	return NULL;
+}
+
+static bool mtk_vdec_get_cap_fmt(struct mtk_vcodec_dec_ctx *ctx, int format_index)
+{
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+	const struct mtk_video_fmt *fmt;
+	struct mtk_q_data *q_data;
+	int num_frame_count = 0, i;
+	bool ret = false;
+
+	fmt = &dec_pdata->vdec_formats[format_index];
+	for (i = 0; i < *dec_pdata->num_formats; i++) {
+		if (dec_pdata->vdec_formats[i].type != MTK_FMT_FRAME)
+			continue;
+
+		num_frame_count++;
+	}
+
+	if (num_frame_count == 1 || (!ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MM21))
+		return true;
+
+	q_data = &ctx->q_data[MTK_Q_DATA_SRC];
+	switch (q_data->fmt->fourcc) {
+	case V4L2_PIX_FMT_H264_SLICE:
+		if (ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MT2110R)
+			ret = true;
+		break;
+	case V4L2_PIX_FMT_VP9_FRAME:
+	case V4L2_PIX_FMT_AV1_FRAME:
+	case V4L2_PIX_FMT_HEVC_SLICE:
+		if (ctx->is_10bit_bitstream && fmt->fourcc == V4L2_PIX_FMT_MT2110T)
+			ret = true;
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+static struct mtk_q_data *mtk_vdec_get_q_data(struct mtk_vcodec_dec_ctx *ctx,
+					      enum v4l2_buf_type type)
+{
+	if (V4L2_TYPE_IS_OUTPUT(type))
+		return &ctx->q_data[MTK_Q_DATA_SRC];
+
+	return &ctx->q_data[MTK_Q_DATA_DST];
+}
+
+static int vidioc_try_decoder_cmd(struct file *file, void *priv,
+				struct v4l2_decoder_cmd *cmd)
+{
+	return v4l2_m2m_ioctl_try_decoder_cmd(file, priv, cmd);
+}
+
+
+static int vidioc_decoder_cmd(struct file *file, void *priv,
+				struct v4l2_decoder_cmd *cmd)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct vb2_queue *src_vq, *dst_vq;
+	int ret;
+
+	ret = vidioc_try_decoder_cmd(file, priv, cmd);
+	if (ret)
+		return ret;
+
+	mtk_v4l2_vdec_dbg(1, ctx, "decoder cmd=%u", cmd->cmd);
+	dst_vq = v4l2_m2m_get_vq(ctx->m2m_ctx,
+				V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	switch (cmd->cmd) {
+	case V4L2_DEC_CMD_STOP:
+		src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx,
+				V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
+		if (!vb2_is_streaming(src_vq)) {
+			mtk_v4l2_vdec_dbg(1, ctx, "Output stream is off. No need to flush.");
+			return 0;
+		}
+		if (!vb2_is_streaming(dst_vq)) {
+			mtk_v4l2_vdec_dbg(1, ctx, "Capture stream is off. No need to flush.");
+			return 0;
+		}
+		v4l2_m2m_buf_queue(ctx->m2m_ctx, &ctx->empty_flush_buf.vb);
+		v4l2_m2m_try_schedule(ctx->m2m_ctx);
+		break;
+
+	case V4L2_DEC_CMD_START:
+		vb2_clear_last_buffer_dequeued(dst_vq);
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+void mtk_vdec_unlock(struct mtk_vcodec_dec_ctx *ctx)
+{
+	mutex_unlock(&ctx->dev->dec_mutex[ctx->hw_id]);
+}
+
+void mtk_vdec_lock(struct mtk_vcodec_dec_ctx *ctx)
+{
+	mutex_lock(&ctx->dev->dec_mutex[ctx->hw_id]);
+}
+
+void mtk_vcodec_dec_release(struct mtk_vcodec_dec_ctx *ctx)
+{
+	vdec_if_deinit(ctx);
+	ctx->state = MTK_STATE_FREE;
+}
+
+void mtk_vcodec_dec_set_default_params(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct mtk_q_data *q_data;
+
+	ctx->m2m_ctx->q_lock = &ctx->dev->dev_mutex;
+	ctx->fh.m2m_ctx = ctx->m2m_ctx;
+	ctx->fh.ctrl_handler = &ctx->ctrl_hdl;
+	INIT_WORK(&ctx->decode_work, ctx->dev->vdec_pdata->worker);
+	ctx->colorspace = V4L2_COLORSPACE_REC709;
+	ctx->ycbcr_enc = V4L2_YCBCR_ENC_DEFAULT;
+	ctx->quantization = V4L2_QUANTIZATION_DEFAULT;
+	ctx->xfer_func = V4L2_XFER_FUNC_DEFAULT;
+
+	q_data = &ctx->q_data[MTK_Q_DATA_SRC];
+	memset(q_data, 0, sizeof(struct mtk_q_data));
+	q_data->visible_width = DFT_CFG_WIDTH;
+	q_data->visible_height = DFT_CFG_HEIGHT;
+	q_data->fmt = ctx->dev->vdec_pdata->default_out_fmt;
+	q_data->field = V4L2_FIELD_NONE;
+
+	q_data->sizeimage[0] = DFT_CFG_WIDTH * DFT_CFG_HEIGHT;
+	q_data->bytesperline[0] = 0;
+
+	q_data = &ctx->q_data[MTK_Q_DATA_DST];
+	memset(q_data, 0, sizeof(struct mtk_q_data));
+	q_data->visible_width = DFT_CFG_WIDTH;
+	q_data->visible_height = DFT_CFG_HEIGHT;
+	q_data->coded_width = DFT_CFG_WIDTH;
+	q_data->coded_height = DFT_CFG_HEIGHT;
+	q_data->fmt = ctx->dev->vdec_pdata->default_cap_fmt;
+	q_data->field = V4L2_FIELD_NONE;
+
+	q_data->sizeimage[0] = q_data->coded_width * q_data->coded_height;
+	q_data->bytesperline[0] = q_data->coded_width;
+	q_data->sizeimage[1] = q_data->sizeimage[0] / 2;
+	q_data->bytesperline[1] = q_data->coded_width;
+}
+
+static int vidioc_vdec_qbuf(struct file *file, void *priv,
+			    struct v4l2_buffer *buf)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+
+	if (ctx->state == MTK_STATE_ABORT) {
+		mtk_v4l2_vdec_err(ctx, "[%d] Call on QBUF after unrecoverable error", ctx->id);
+		return -EIO;
+	}
+
+	return v4l2_m2m_qbuf(file, ctx->m2m_ctx, buf);
+}
+
+static int vidioc_vdec_dqbuf(struct file *file, void *priv,
+			     struct v4l2_buffer *buf)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+
+	if (ctx->state == MTK_STATE_ABORT) {
+		mtk_v4l2_vdec_err(ctx, "[%d] Call on DQBUF after unrecoverable error", ctx->id);
+		return -EIO;
+	}
+
+	return v4l2_m2m_dqbuf(file, ctx->m2m_ctx, buf);
+}
+
+static int mtk_vcodec_dec_get_chip_name(void *priv)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct device *dev = &ctx->dev->plat_dev->dev;
+
+	if (of_device_is_compatible(dev->of_node, "mediatek,mt8173-vcodec-dec"))
+		return 8173;
+	else if (of_device_is_compatible(dev->of_node, "mediatek,mt8183-vcodec-dec"))
+		return 8183;
+	else if (of_device_is_compatible(dev->of_node, "mediatek,mt8192-vcodec-dec"))
+		return 8192;
+	else if (of_device_is_compatible(dev->of_node, "mediatek,mt8195-vcodec-dec"))
+		return 8195;
+	else if (of_device_is_compatible(dev->of_node, "mediatek,mt8186-vcodec-dec"))
+		return 8186;
+	else if (of_device_is_compatible(dev->of_node, "mediatek,mt8188-vcodec-dec"))
+		return 8188;
+	else
+		return 8173;
+}
+
+static int vidioc_vdec_querycap(struct file *file, void *priv,
+				struct v4l2_capability *cap)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct device *dev = &ctx->dev->plat_dev->dev;
+	int platform_name = mtk_vcodec_dec_get_chip_name(priv);
+
+	strscpy(cap->driver, dev->driver->name, sizeof(cap->driver));
+	snprintf(cap->card, sizeof(cap->card), "MT%d video decoder", platform_name);
+
+	return 0;
+}
+
+static int vidioc_vdec_subscribe_evt(struct v4l2_fh *fh,
+				     const struct v4l2_event_subscription *sub)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(fh);
+
+	if (ctx->dev->vdec_pdata->uses_stateless_api)
+		return v4l2_ctrl_subscribe_event(fh, sub);
+
+	switch (sub->type) {
+	case V4L2_EVENT_EOS:
+		return v4l2_event_subscribe(fh, sub, 2, NULL);
+	case V4L2_EVENT_SOURCE_CHANGE:
+		return v4l2_src_change_event_subscribe(fh, sub);
+	default:
+		return v4l2_ctrl_subscribe_event(fh, sub);
+	}
+}
+
+static int vidioc_try_fmt(struct mtk_vcodec_dec_ctx *ctx, struct v4l2_format *f,
+			  const struct mtk_video_fmt *fmt)
+{
+	struct v4l2_pix_format_mplane *pix_fmt_mp = &f->fmt.pix_mp;
+	const struct v4l2_frmsize_stepwise *frmsize;
+
+	pix_fmt_mp->field = V4L2_FIELD_NONE;
+
+	/* Always apply frame size constraints from the coded side */
+	if (V4L2_TYPE_IS_OUTPUT(f->type))
+		frmsize = &fmt->frmsize;
+	else
+		frmsize = &ctx->q_data[MTK_Q_DATA_SRC].fmt->frmsize;
+
+	pix_fmt_mp->width = clamp(pix_fmt_mp->width, MTK_VDEC_MIN_W, frmsize->max_width);
+	pix_fmt_mp->height = clamp(pix_fmt_mp->height, MTK_VDEC_MIN_H, frmsize->max_height);
+
+	if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+		pix_fmt_mp->num_planes = 1;
+		pix_fmt_mp->plane_fmt[0].bytesperline = 0;
+	} else if (f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
+		int tmp_w, tmp_h;
+
+		/*
+		 * Find next closer width align 64, heign align 64, size align
+		 * 64 rectangle
+		 * Note: This only get default value, the real HW needed value
+		 *       only available when ctx in MTK_STATE_HEADER state
+		 */
+		tmp_w = pix_fmt_mp->width;
+		tmp_h = pix_fmt_mp->height;
+		v4l_bound_align_image(&pix_fmt_mp->width, MTK_VDEC_MIN_W, frmsize->max_width, 6,
+				      &pix_fmt_mp->height, MTK_VDEC_MIN_H, frmsize->max_height, 6,
+				      9);
+
+		if (pix_fmt_mp->width < tmp_w &&
+		    (pix_fmt_mp->width + 64) <= frmsize->max_width)
+			pix_fmt_mp->width += 64;
+		if (pix_fmt_mp->height < tmp_h &&
+		    (pix_fmt_mp->height + 64) <= frmsize->max_height)
+			pix_fmt_mp->height += 64;
+
+		mtk_v4l2_vdec_dbg(0, ctx,
+				  "before resize wxh=%dx%d, after resize wxh=%dx%d, sizeimage=%d",
+				  tmp_w, tmp_h, pix_fmt_mp->width, pix_fmt_mp->height,
+				  pix_fmt_mp->width * pix_fmt_mp->height);
+
+		pix_fmt_mp->num_planes = fmt->num_planes;
+		pix_fmt_mp->plane_fmt[0].sizeimage =
+				pix_fmt_mp->width * pix_fmt_mp->height;
+		pix_fmt_mp->plane_fmt[0].bytesperline = pix_fmt_mp->width;
+
+		if (pix_fmt_mp->num_planes == 2) {
+			pix_fmt_mp->plane_fmt[1].sizeimage =
+				(pix_fmt_mp->width * pix_fmt_mp->height) / 2;
+			pix_fmt_mp->plane_fmt[1].bytesperline =
+				pix_fmt_mp->width;
+		}
+	}
+
+	pix_fmt_mp->flags = 0;
+	return 0;
+}
+
+static int vidioc_try_fmt_vid_cap_mplane(struct file *file, void *priv,
+				struct v4l2_format *f)
+{
+	const struct mtk_video_fmt *fmt;
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+
+	fmt = mtk_vdec_find_format(f, dec_pdata);
+	if (!fmt) {
+		f->fmt.pix.pixelformat =
+			ctx->q_data[MTK_Q_DATA_DST].fmt->fourcc;
+		fmt = mtk_vdec_find_format(f, dec_pdata);
+	}
+
+	return vidioc_try_fmt(ctx, f, fmt);
+}
+
+static int vidioc_try_fmt_vid_out_mplane(struct file *file, void *priv,
+				struct v4l2_format *f)
+{
+	struct v4l2_pix_format_mplane *pix_fmt_mp = &f->fmt.pix_mp;
+	const struct mtk_video_fmt *fmt;
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+
+	fmt = mtk_vdec_find_format(f, dec_pdata);
+	if (!fmt) {
+		f->fmt.pix.pixelformat =
+			ctx->q_data[MTK_Q_DATA_SRC].fmt->fourcc;
+		fmt = mtk_vdec_find_format(f, dec_pdata);
+	}
+
+	if (pix_fmt_mp->plane_fmt[0].sizeimage == 0) {
+		mtk_v4l2_vdec_err(ctx, "sizeimage of output format must be given");
+		return -EINVAL;
+	}
+
+	return vidioc_try_fmt(ctx, f, fmt);
+}
+
+static int vidioc_vdec_g_selection(struct file *file, void *priv,
+			struct v4l2_selection *s)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct mtk_q_data *q_data;
+
+	if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE)
+		return -EINVAL;
+
+	q_data = &ctx->q_data[MTK_Q_DATA_DST];
+
+	switch (s->target) {
+	case V4L2_SEL_TGT_COMPOSE_DEFAULT:
+		s->r.left = 0;
+		s->r.top = 0;
+		s->r.width = ctx->picinfo.pic_w;
+		s->r.height = ctx->picinfo.pic_h;
+		break;
+	case V4L2_SEL_TGT_COMPOSE_BOUNDS:
+		s->r.left = 0;
+		s->r.top = 0;
+		s->r.width = ctx->picinfo.buf_w;
+		s->r.height = ctx->picinfo.buf_h;
+		break;
+	case V4L2_SEL_TGT_COMPOSE:
+		if (vdec_if_get_param(ctx, GET_PARAM_CROP_INFO, &(s->r))) {
+			/* set to default value if header info not ready yet*/
+			s->r.left = 0;
+			s->r.top = 0;
+			s->r.width = q_data->visible_width;
+			s->r.height = q_data->visible_height;
+		}
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	if (ctx->state < MTK_STATE_HEADER) {
+		/* set to default value if header info not ready yet*/
+		s->r.left = 0;
+		s->r.top = 0;
+		s->r.width = q_data->visible_width;
+		s->r.height = q_data->visible_height;
+		return 0;
+	}
+
+	return 0;
+}
+
+static int vidioc_vdec_s_selection(struct file *file, void *priv,
+				struct v4l2_selection *s)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+
+	if (s->type != V4L2_BUF_TYPE_VIDEO_CAPTURE)
+		return -EINVAL;
+
+	switch (s->target) {
+	case V4L2_SEL_TGT_COMPOSE:
+		s->r.left = 0;
+		s->r.top = 0;
+		s->r.width = ctx->picinfo.pic_w;
+		s->r.height = ctx->picinfo.pic_h;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int vidioc_vdec_s_fmt(struct file *file, void *priv,
+			     struct v4l2_format *f)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct v4l2_pix_format_mplane *pix_mp;
+	struct mtk_q_data *q_data;
+	int ret = 0;
+	const struct mtk_video_fmt *fmt;
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id);
+
+	q_data = mtk_vdec_get_q_data(ctx, f->type);
+	if (!q_data)
+		return -EINVAL;
+
+	pix_mp = &f->fmt.pix_mp;
+	/*
+	 * Setting OUTPUT format after OUTPUT buffers are allocated is invalid
+	 * if using the stateful API.
+	 */
+	if (!dec_pdata->uses_stateless_api &&
+	    f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE &&
+	    vb2_is_busy(&ctx->m2m_ctx->out_q_ctx.q)) {
+		mtk_v4l2_vdec_err(ctx, "out_q_ctx buffers already requested");
+		ret = -EBUSY;
+	}
+
+	/*
+	 * Setting CAPTURE format after CAPTURE buffers are allocated is
+	 * invalid.
+	 */
+	if ((f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) &&
+	    vb2_is_busy(&ctx->m2m_ctx->cap_q_ctx.q)) {
+		mtk_v4l2_vdec_err(ctx, "cap_q_ctx buffers already requested");
+		ret = -EBUSY;
+	}
+
+	fmt = mtk_vdec_find_format(f, dec_pdata);
+	if (fmt == NULL) {
+		if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+			f->fmt.pix.pixelformat =
+				dec_pdata->default_out_fmt->fourcc;
+			fmt = mtk_vdec_find_format(f, dec_pdata);
+		} else if (f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
+			f->fmt.pix.pixelformat =
+				dec_pdata->default_cap_fmt->fourcc;
+			fmt = mtk_vdec_find_format(f, dec_pdata);
+		}
+	}
+	if (fmt == NULL)
+		return -EINVAL;
+
+	q_data->fmt = fmt;
+	vidioc_try_fmt(ctx, f, q_data->fmt);
+	if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+		q_data->sizeimage[0] = pix_mp->plane_fmt[0].sizeimage;
+		q_data->coded_width = pix_mp->width;
+		q_data->coded_height = pix_mp->height;
+
+		ctx->colorspace = pix_mp->colorspace;
+		ctx->ycbcr_enc = pix_mp->ycbcr_enc;
+		ctx->quantization = pix_mp->quantization;
+		ctx->xfer_func = pix_mp->xfer_func;
+
+		ctx->current_codec = fmt->fourcc;
+		if (ctx->state == MTK_STATE_FREE) {
+			ret = vdec_if_init(ctx, q_data->fmt->fourcc);
+			if (ret) {
+				mtk_v4l2_vdec_err(ctx, "[%d]: vdec_if_init() fail ret=%d",
+						  ctx->id, ret);
+				return -EINVAL;
+			}
+			ctx->state = MTK_STATE_INIT;
+		}
+	} else {
+		ctx->capture_fourcc = fmt->fourcc;
+	}
+
+	/*
+	 * If using the stateless API, S_FMT should have the effect of setting
+	 * the CAPTURE queue resolution no matter which queue it was called on.
+	 */
+	if (dec_pdata->uses_stateless_api) {
+		ctx->picinfo.pic_w = pix_mp->width;
+		ctx->picinfo.pic_h = pix_mp->height;
+
+		/*
+		 * If get pic info fail, need to use the default pic info params, or
+		 * v4l2-compliance will fail
+		 */
+		ret = vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo);
+		if (ret) {
+			mtk_v4l2_vdec_err(ctx, "[%d]Error!! Get GET_PARAM_PICTURE_INFO Fail",
+					  ctx->id);
+		}
+
+		ctx->last_decoded_picinfo = ctx->picinfo;
+
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1) {
+			ctx->q_data[MTK_Q_DATA_DST].sizeimage[0] =
+				ctx->picinfo.fb_sz[0] +
+				ctx->picinfo.fb_sz[1];
+			ctx->q_data[MTK_Q_DATA_DST].bytesperline[0] =
+				ctx->picinfo.buf_w;
+		} else {
+			ctx->q_data[MTK_Q_DATA_DST].sizeimage[0] =
+				ctx->picinfo.fb_sz[0];
+			ctx->q_data[MTK_Q_DATA_DST].bytesperline[0] =
+				ctx->picinfo.buf_w;
+			ctx->q_data[MTK_Q_DATA_DST].sizeimage[1] =
+				ctx->picinfo.fb_sz[1];
+			ctx->q_data[MTK_Q_DATA_DST].bytesperline[1] =
+				ctx->picinfo.buf_w;
+		}
+
+		ctx->q_data[MTK_Q_DATA_DST].coded_width = ctx->picinfo.buf_w;
+		ctx->q_data[MTK_Q_DATA_DST].coded_height = ctx->picinfo.buf_h;
+		mtk_v4l2_vdec_dbg(2, ctx,
+				  "[%d] init() plane:%d wxh=%dx%d pic wxh=%dx%d sz=0x%x_0x%x",
+				  ctx->id, pix_mp->num_planes,
+				  ctx->picinfo.buf_w, ctx->picinfo.buf_h,
+				  ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+				  ctx->q_data[MTK_Q_DATA_DST].sizeimage[0],
+				  ctx->q_data[MTK_Q_DATA_DST].sizeimage[1]);
+	}
+	return 0;
+}
+
+static int vidioc_enum_framesizes(struct file *file, void *priv,
+				struct v4l2_frmsizeenum *fsize)
+{
+	int i = 0;
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+
+	if (fsize->index != 0)
+		return -EINVAL;
+
+	for (i = 0; i < *dec_pdata->num_formats; i++) {
+		if (fsize->pixel_format != dec_pdata->vdec_formats[i].fourcc)
+			continue;
+
+		/* Only coded formats have frame sizes set */
+		if (!dec_pdata->vdec_formats[i].frmsize.max_width)
+			return -ENOTTY;
+
+		fsize->type = V4L2_FRMSIZE_TYPE_STEPWISE;
+		fsize->stepwise = dec_pdata->vdec_formats[i].frmsize;
+
+		mtk_v4l2_vdec_dbg(1, ctx, "%x, %d %d %d %d %d %d",
+				  ctx->dev->dec_capability, fsize->stepwise.min_width,
+				  fsize->stepwise.max_width, fsize->stepwise.step_width,
+				  fsize->stepwise.min_height, fsize->stepwise.max_height,
+				  fsize->stepwise.step_height);
+
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+static int vidioc_enum_fmt(struct v4l2_fmtdesc *f, void *priv,
+			   bool output_queue)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+	const struct mtk_video_fmt *fmt;
+	int i, j = 0;
+
+	for (i = 0; i < *dec_pdata->num_formats; i++) {
+		if (output_queue &&
+		    dec_pdata->vdec_formats[i].type != MTK_FMT_DEC)
+			continue;
+		if (!output_queue &&
+		    dec_pdata->vdec_formats[i].type != MTK_FMT_FRAME)
+			continue;
+
+		if (!output_queue && !mtk_vdec_get_cap_fmt(ctx, i))
+			continue;
+
+		if (j == f->index)
+			break;
+		++j;
+	}
+
+	if (i == *dec_pdata->num_formats)
+		return -EINVAL;
+
+	fmt = &dec_pdata->vdec_formats[i];
+	f->pixelformat = fmt->fourcc;
+	f->flags = fmt->flags;
+
+	return 0;
+}
+
+static int vidioc_vdec_enum_fmt_vid_cap(struct file *file, void *priv,
+					struct v4l2_fmtdesc *f)
+{
+	return vidioc_enum_fmt(f, priv, false);
+}
+
+static int vidioc_vdec_enum_fmt_vid_out(struct file *file, void *priv,
+					struct v4l2_fmtdesc *f)
+{
+	return vidioc_enum_fmt(f, priv, true);
+}
+
+static int vidioc_vdec_g_fmt(struct file *file, void *priv,
+			     struct v4l2_format *f)
+{
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(priv);
+	struct v4l2_pix_format_mplane *pix_mp = &f->fmt.pix_mp;
+	struct vb2_queue *vq;
+	struct mtk_q_data *q_data;
+
+	vq = v4l2_m2m_get_vq(ctx->m2m_ctx, f->type);
+	if (!vq) {
+		mtk_v4l2_vdec_err(ctx, "no vb2 queue for type=%d", f->type);
+		return -EINVAL;
+	}
+
+	q_data = mtk_vdec_get_q_data(ctx, f->type);
+
+	pix_mp->field = V4L2_FIELD_NONE;
+	pix_mp->colorspace = ctx->colorspace;
+	pix_mp->ycbcr_enc = ctx->ycbcr_enc;
+	pix_mp->quantization = ctx->quantization;
+	pix_mp->xfer_func = ctx->xfer_func;
+
+	if ((f->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) &&
+	    (ctx->state >= MTK_STATE_HEADER)) {
+		/* Until STREAMOFF is called on the CAPTURE queue
+		 * (acknowledging the event), the driver operates as if
+		 * the resolution hasn't changed yet.
+		 * So we just return picinfo yet, and update picinfo in
+		 * stop_streaming hook function
+		 */
+		q_data->sizeimage[0] = ctx->picinfo.fb_sz[0];
+		q_data->sizeimage[1] = ctx->picinfo.fb_sz[1];
+		q_data->bytesperline[0] = ctx->last_decoded_picinfo.buf_w;
+		q_data->bytesperline[1] = ctx->last_decoded_picinfo.buf_w;
+		q_data->coded_width = ctx->picinfo.buf_w;
+		q_data->coded_height = ctx->picinfo.buf_h;
+		ctx->last_decoded_picinfo.cap_fourcc = q_data->fmt->fourcc;
+
+		/*
+		 * Width and height are set to the dimensions
+		 * of the movie, the buffer is bigger and
+		 * further processing stages should crop to this
+		 * rectangle.
+		 */
+		pix_mp->width = q_data->coded_width;
+		pix_mp->height = q_data->coded_height;
+
+		/*
+		 * Set pixelformat to the format in which mt vcodec
+		 * outputs the decoded frame
+		 */
+		pix_mp->num_planes = q_data->fmt->num_planes;
+		pix_mp->pixelformat = q_data->fmt->fourcc;
+		pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0];
+		pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0];
+		pix_mp->plane_fmt[1].bytesperline = q_data->bytesperline[1];
+		pix_mp->plane_fmt[1].sizeimage = q_data->sizeimage[1];
+
+	} else if (f->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+		/*
+		 * This is run on OUTPUT
+		 * The buffer contains compressed image
+		 * so width and height have no meaning.
+		 * Assign value here to pass v4l2-compliance test
+		 */
+		pix_mp->width = q_data->visible_width;
+		pix_mp->height = q_data->visible_height;
+		pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0];
+		pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0];
+		pix_mp->pixelformat = q_data->fmt->fourcc;
+		pix_mp->num_planes = q_data->fmt->num_planes;
+	} else {
+		pix_mp->width = q_data->coded_width;
+		pix_mp->height = q_data->coded_height;
+		pix_mp->num_planes = q_data->fmt->num_planes;
+		pix_mp->pixelformat = q_data->fmt->fourcc;
+		pix_mp->plane_fmt[0].bytesperline = q_data->bytesperline[0];
+		pix_mp->plane_fmt[0].sizeimage = q_data->sizeimage[0];
+		pix_mp->plane_fmt[1].bytesperline = q_data->bytesperline[1];
+		pix_mp->plane_fmt[1].sizeimage = q_data->sizeimage[1];
+
+		mtk_v4l2_vdec_dbg(1, ctx, "[%d] type=%d state=%d Format information not ready!",
+				  ctx->id, f->type, ctx->state);
+	}
+
+	return 0;
+}
+
+int vb2ops_vdec_queue_setup(struct vb2_queue *vq, unsigned int *nbuffers,
+			    unsigned int *nplanes, unsigned int sizes[],
+			    struct device *alloc_devs[])
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vq);
+	struct mtk_q_data *q_data;
+	unsigned int i;
+
+	q_data = mtk_vdec_get_q_data(ctx, vq->type);
+
+	if (q_data == NULL) {
+		mtk_v4l2_vdec_err(ctx, "vq->type=%d err\n", vq->type);
+		return -EINVAL;
+	}
+
+	if (*nplanes) {
+		if (vq->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
+			if (*nplanes != q_data->fmt->num_planes)
+				return -EINVAL;
+		} else {
+			if (*nplanes != 1)
+				return -EINVAL;
+		}
+		for (i = 0; i < *nplanes; i++) {
+			if (sizes[i] < q_data->sizeimage[i])
+				return -EINVAL;
+		}
+	} else {
+		if (vq->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE)
+			*nplanes = q_data->fmt->num_planes;
+		else
+			*nplanes = 1;
+
+		for (i = 0; i < *nplanes; i++)
+			sizes[i] = q_data->sizeimage[i];
+	}
+
+	mtk_v4l2_vdec_dbg(1, ctx,
+			  "[%d]\t type = %d, get %d plane(s), %d buffer(s) of size 0x%x 0x%x ",
+			  ctx->id, vq->type, *nplanes, *nbuffers, sizes[0], sizes[1]);
+
+	return 0;
+}
+
+int vb2ops_vdec_buf_prepare(struct vb2_buffer *vb)
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+	struct mtk_q_data *q_data;
+	int i;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d",
+			  ctx->id, vb->vb2_queue->type, vb->index);
+
+	q_data = mtk_vdec_get_q_data(ctx, vb->vb2_queue->type);
+
+	for (i = 0; i < q_data->fmt->num_planes; i++) {
+		if (vb2_plane_size(vb, i) < q_data->sizeimage[i]) {
+			mtk_v4l2_vdec_err(ctx, "data will not fit into plane %d (%lu < %d)",
+					  i, vb2_plane_size(vb, i), q_data->sizeimage[i]);
+			return -EINVAL;
+		}
+		if (!V4L2_TYPE_IS_OUTPUT(vb->type))
+			vb2_set_plane_payload(vb, i, q_data->sizeimage[i]);
+	}
+
+	return 0;
+}
+
+void vb2ops_vdec_buf_finish(struct vb2_buffer *vb)
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	struct mtk_video_dec_buf *buf;
+	bool buf_error;
+
+	vb2_v4l2 = container_of(vb, struct vb2_v4l2_buffer, vb2_buf);
+	buf = container_of(vb2_v4l2, struct mtk_video_dec_buf, m2m_buf.vb);
+	mutex_lock(&ctx->lock);
+	if (vb->vb2_queue->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
+		buf->queued_in_v4l2 = false;
+		buf->queued_in_vb2 = false;
+	}
+	buf_error = buf->error;
+	mutex_unlock(&ctx->lock);
+
+	if (buf_error) {
+		mtk_v4l2_vdec_err(ctx, "Unrecoverable error on buffer.");
+		ctx->state = MTK_STATE_ABORT;
+	}
+}
+
+int vb2ops_vdec_buf_init(struct vb2_buffer *vb)
+{
+	struct vb2_v4l2_buffer *vb2_v4l2 = container_of(vb,
+					struct vb2_v4l2_buffer, vb2_buf);
+	struct mtk_video_dec_buf *buf = container_of(vb2_v4l2,
+					struct mtk_video_dec_buf, m2m_buf.vb);
+
+	if (vb->vb2_queue->type == V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE) {
+		buf->used = false;
+		buf->queued_in_v4l2 = false;
+	}
+
+	return 0;
+}
+
+int vb2ops_vdec_start_streaming(struct vb2_queue *q, unsigned int count)
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(q);
+
+	if (ctx->state == MTK_STATE_FLUSH)
+		ctx->state = MTK_STATE_HEADER;
+
+	return 0;
+}
+
+void vb2ops_vdec_stop_streaming(struct vb2_queue *q)
+{
+	struct vb2_v4l2_buffer *src_buf = NULL, *dst_buf = NULL;
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(q);
+	int ret;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) state=(%x) ctx->decoded_frame_cnt=%d",
+			  ctx->id, q->type, ctx->state, ctx->decoded_frame_cnt);
+
+	if (q->type == V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+		while ((src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx))) {
+			if (src_buf != &ctx->empty_flush_buf.vb) {
+				struct media_request *req =
+					src_buf->vb2_buf.req_obj.req;
+				v4l2_m2m_buf_done(src_buf,
+						VB2_BUF_STATE_ERROR);
+				if (req)
+					v4l2_ctrl_request_complete(req, &ctx->ctrl_hdl);
+			}
+		}
+		return;
+	}
+
+	if (ctx->state >= MTK_STATE_HEADER) {
+
+		/* Until STREAMOFF is called on the CAPTURE queue
+		 * (acknowledging the event), the driver operates
+		 * as if the resolution hasn't changed yet, i.e.
+		 * VIDIOC_G_FMT< etc. return previous resolution.
+		 * So we update picinfo here
+		 */
+		ctx->picinfo = ctx->last_decoded_picinfo;
+
+		mtk_v4l2_vdec_dbg(2, ctx,
+				  "[%d]-> new(%d,%d), old(%d,%d), real(%d,%d)",
+				  ctx->id, ctx->last_decoded_picinfo.pic_w,
+				  ctx->last_decoded_picinfo.pic_h,
+				  ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+				  ctx->last_decoded_picinfo.buf_w,
+				  ctx->last_decoded_picinfo.buf_h);
+
+		ret = ctx->dev->vdec_pdata->flush_decoder(ctx);
+		if (ret)
+			mtk_v4l2_vdec_err(ctx, "DecodeFinal failed, ret=%d", ret);
+	}
+	ctx->state = MTK_STATE_FLUSH;
+
+	while ((dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx))) {
+		vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0);
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2)
+			vb2_set_plane_payload(&dst_buf->vb2_buf, 1, 0);
+		v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_ERROR);
+	}
+
+}
+
+static void m2mops_vdec_device_run(void *priv)
+{
+	struct mtk_vcodec_dec_ctx *ctx = priv;
+	struct mtk_vcodec_dec_dev *dev = ctx->dev;
+
+	queue_work(dev->decode_workqueue, &ctx->decode_work);
+}
+
+static int m2mops_vdec_job_ready(void *m2m_priv)
+{
+	struct mtk_vcodec_dec_ctx *ctx = m2m_priv;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id);
+
+	if (ctx->state == MTK_STATE_ABORT)
+		return 0;
+
+	if ((ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w) ||
+	    (ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h))
+		return 0;
+
+	if (ctx->state != MTK_STATE_HEADER)
+		return 0;
+
+	return 1;
+}
+
+static void m2mops_vdec_job_abort(void *priv)
+{
+	struct mtk_vcodec_dec_ctx *ctx = priv;
+
+	ctx->state = MTK_STATE_ABORT;
+}
+
+const struct v4l2_m2m_ops mtk_vdec_m2m_ops = {
+	.device_run	= m2mops_vdec_device_run,
+	.job_ready	= m2mops_vdec_job_ready,
+	.job_abort	= m2mops_vdec_job_abort,
+};
+
+const struct v4l2_ioctl_ops mtk_vdec_ioctl_ops = {
+	.vidioc_streamon	= v4l2_m2m_ioctl_streamon,
+	.vidioc_streamoff	= v4l2_m2m_ioctl_streamoff,
+	.vidioc_reqbufs		= v4l2_m2m_ioctl_reqbufs,
+	.vidioc_querybuf	= v4l2_m2m_ioctl_querybuf,
+	.vidioc_expbuf		= v4l2_m2m_ioctl_expbuf,
+
+	.vidioc_qbuf		= vidioc_vdec_qbuf,
+	.vidioc_dqbuf		= vidioc_vdec_dqbuf,
+
+	.vidioc_try_fmt_vid_cap_mplane	= vidioc_try_fmt_vid_cap_mplane,
+	.vidioc_try_fmt_vid_out_mplane	= vidioc_try_fmt_vid_out_mplane,
+
+	.vidioc_s_fmt_vid_cap_mplane	= vidioc_vdec_s_fmt,
+	.vidioc_s_fmt_vid_out_mplane	= vidioc_vdec_s_fmt,
+	.vidioc_g_fmt_vid_cap_mplane	= vidioc_vdec_g_fmt,
+	.vidioc_g_fmt_vid_out_mplane	= vidioc_vdec_g_fmt,
+
+	.vidioc_create_bufs		= v4l2_m2m_ioctl_create_bufs,
+
+	.vidioc_enum_fmt_vid_cap	= vidioc_vdec_enum_fmt_vid_cap,
+	.vidioc_enum_fmt_vid_out	= vidioc_vdec_enum_fmt_vid_out,
+	.vidioc_enum_framesizes	= vidioc_enum_framesizes,
+
+	.vidioc_querycap		= vidioc_vdec_querycap,
+	.vidioc_subscribe_event		= vidioc_vdec_subscribe_evt,
+	.vidioc_unsubscribe_event	= v4l2_event_unsubscribe,
+	.vidioc_g_selection             = vidioc_vdec_g_selection,
+	.vidioc_s_selection             = vidioc_vdec_s_selection,
+
+	.vidioc_decoder_cmd = vidioc_decoder_cmd,
+	.vidioc_try_decoder_cmd = vidioc_try_decoder_cmd,
+};
+
+int mtk_vcodec_dec_queue_init(void *priv, struct vb2_queue *src_vq,
+			   struct vb2_queue *dst_vq)
+{
+	struct mtk_vcodec_dec_ctx *ctx = priv;
+	int ret = 0;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id);
+
+	src_vq->type		= V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	src_vq->io_modes	= VB2_DMABUF | VB2_MMAP;
+	src_vq->drv_priv	= ctx;
+	src_vq->buf_struct_size = sizeof(struct mtk_video_dec_buf);
+	src_vq->ops		= ctx->dev->vdec_pdata->vdec_vb2_ops;
+	src_vq->mem_ops		= &vb2_dma_contig_memops;
+	src_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
+	src_vq->lock		= &ctx->dev->dev_mutex;
+	src_vq->dev             = &ctx->dev->plat_dev->dev;
+	src_vq->allow_cache_hints = 1;
+
+	ret = vb2_queue_init(src_vq);
+	if (ret) {
+		mtk_v4l2_vdec_err(ctx, "Failed to initialize videobuf2 queue(output)");
+		return ret;
+	}
+	dst_vq->type		= V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
+	dst_vq->io_modes	= VB2_DMABUF | VB2_MMAP;
+	dst_vq->drv_priv	= ctx;
+	dst_vq->buf_struct_size = sizeof(struct mtk_video_dec_buf);
+	dst_vq->ops		= ctx->dev->vdec_pdata->vdec_vb2_ops;
+	dst_vq->mem_ops		= &vb2_dma_contig_memops;
+	dst_vq->timestamp_flags = V4L2_BUF_FLAG_TIMESTAMP_COPY;
+	dst_vq->lock		= &ctx->dev->dev_mutex;
+	dst_vq->dev             = &ctx->dev->plat_dev->dev;
+	dst_vq->allow_cache_hints = 1;
+
+	ret = vb2_queue_init(dst_vq);
+	if (ret)
+		mtk_v4l2_vdec_err(ctx, "Failed to initialize videobuf2 queue(capture)");
+
+	return ret;
+}
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h
new file mode 100644
index 0000000000..ece27c880e
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ *         Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#ifndef _MTK_VCODEC_DEC_H_
+#define _MTK_VCODEC_DEC_H_
+
+#include <media/videobuf2-core.h>
+#include <media/v4l2-mem2mem.h>
+
+#include "mtk_vcodec_dec_drv.h"
+
+#define VCODEC_DEC_ALIGNED_64 64
+#define VCODEC_CAPABILITY_4K_DISABLED	0x10
+#define VCODEC_DEC_4K_CODED_WIDTH	4096U
+#define VCODEC_DEC_4K_CODED_HEIGHT	2304U
+#define MTK_VDEC_MAX_W	2048U
+#define MTK_VDEC_MAX_H	1088U
+#define MTK_VDEC_MIN_W	64U
+#define MTK_VDEC_MIN_H	64U
+
+#define MTK_VDEC_IRQ_STATUS_DEC_SUCCESS        0x10000
+
+/**
+ * struct vdec_fb  - decoder frame buffer
+ * @base_y	: Y plane memory info
+ * @base_c	: C plane memory info
+ * @status      : frame buffer status (vdec_fb_status)
+ */
+struct vdec_fb {
+	struct mtk_vcodec_mem	base_y;
+	struct mtk_vcodec_mem	base_c;
+	unsigned int	status;
+};
+
+/**
+ * struct mtk_video_dec_buf - Private data related to each VB2 buffer.
+ * @m2m_buf:	M2M buffer
+ * @list:	link list
+ * @used:	Capture buffer contain decoded frame data and keep in
+ *			codec data structure
+ * @queued_in_vb2:	Capture buffer is queue in vb2
+ * @queued_in_v4l2:	Capture buffer is in v4l2 driver, but not in vb2
+ *			queue yet
+ * @error:		An unrecoverable error occurs on this buffer.
+ * @frame_buffer:	Decode status, and buffer information of Capture buffer
+ * @bs_buffer:	Output buffer info
+ *
+ * Note : These status information help us track and debug buffer state
+ */
+struct mtk_video_dec_buf {
+	struct v4l2_m2m_buffer	m2m_buf;
+
+	bool	used;
+	bool	queued_in_vb2;
+	bool	queued_in_v4l2;
+	bool	error;
+
+	union {
+		struct vdec_fb	frame_buffer;
+		struct mtk_vcodec_mem	bs_buffer;
+	};
+};
+
+extern const struct v4l2_ioctl_ops mtk_vdec_ioctl_ops;
+extern const struct v4l2_m2m_ops mtk_vdec_m2m_ops;
+extern const struct media_device_ops mtk_vcodec_media_ops;
+extern const struct mtk_vcodec_dec_pdata mtk_vdec_8173_pdata;
+extern const struct mtk_vcodec_dec_pdata mtk_vdec_8183_pdata;
+extern const struct mtk_vcodec_dec_pdata mtk_lat_sig_core_pdata;
+extern const struct mtk_vcodec_dec_pdata mtk_vdec_single_core_pdata;
+
+
+/*
+ * mtk_vdec_lock/mtk_vdec_unlock are for ctx instance to
+ * get/release lock before/after access decoder hw.
+ * mtk_vdec_lock get decoder hw lock and set curr_ctx
+ * to ctx instance that get lock
+ */
+void mtk_vdec_unlock(struct mtk_vcodec_dec_ctx *ctx);
+void mtk_vdec_lock(struct mtk_vcodec_dec_ctx *ctx);
+int mtk_vcodec_dec_queue_init(void *priv, struct vb2_queue *src_vq,
+			   struct vb2_queue *dst_vq);
+void mtk_vcodec_dec_set_default_params(struct mtk_vcodec_dec_ctx *ctx);
+void mtk_vcodec_dec_release(struct mtk_vcodec_dec_ctx *ctx);
+
+/*
+ * VB2 ops
+ */
+int vb2ops_vdec_queue_setup(struct vb2_queue *vq, unsigned int *nbuffers,
+			    unsigned int *nplanes, unsigned int sizes[],
+			    struct device *alloc_devs[]);
+int vb2ops_vdec_buf_prepare(struct vb2_buffer *vb);
+void vb2ops_vdec_buf_finish(struct vb2_buffer *vb);
+int vb2ops_vdec_buf_init(struct vb2_buffer *vb);
+int vb2ops_vdec_start_streaming(struct vb2_queue *q, unsigned int count);
+void vb2ops_vdec_stop_streaming(struct vb2_queue *q);
+
+
+#endif /* _MTK_VCODEC_DEC_H_ */
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c
new file mode 100644
index 0000000000..0a89ce452a
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c
@@ -0,0 +1,573 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ *         Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#include <linux/bitfield.h>
+#include <linux/slab.h>
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+#include <media/v4l2-event.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+#include <media/v4l2-device.h>
+
+#include "mtk_vcodec_dec.h"
+#include "mtk_vcodec_dec_hw.h"
+#include "mtk_vcodec_dec_pm.h"
+#include "../common/mtk_vcodec_intr.h"
+
+static int mtk_vcodec_get_hw_count(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_dec_dev *dev)
+{
+	switch (dev->vdec_pdata->hw_arch) {
+	case MTK_VDEC_PURE_SINGLE_CORE:
+		return MTK_VDEC_ONE_CORE;
+	case MTK_VDEC_LAT_SINGLE_CORE:
+		return MTK_VDEC_ONE_LAT_ONE_CORE;
+	default:
+		mtk_v4l2_vdec_err(ctx, "hw arch %d not supported", dev->vdec_pdata->hw_arch);
+		return MTK_VDEC_NO_HW;
+	}
+}
+
+static bool mtk_vcodec_is_hw_active(struct mtk_vcodec_dec_dev *dev)
+{
+	u32 cg_status;
+
+	if (dev->vdecsys_regmap)
+		return !regmap_test_bits(dev->vdecsys_regmap, VDEC_HW_ACTIVE_ADDR,
+					 VDEC_HW_ACTIVE_MASK);
+
+	cg_status = readl(dev->reg_base[VDEC_SYS] + VDEC_HW_ACTIVE_ADDR);
+	return !FIELD_GET(VDEC_HW_ACTIVE_MASK, cg_status);
+}
+
+static irqreturn_t mtk_vcodec_dec_irq_handler(int irq, void *priv)
+{
+	struct mtk_vcodec_dec_dev *dev = priv;
+	struct mtk_vcodec_dec_ctx *ctx;
+	unsigned int dec_done_status = 0;
+	void __iomem *vdec_misc_addr = dev->reg_base[VDEC_MISC] +
+					VDEC_IRQ_CFG_REG;
+
+	ctx = mtk_vcodec_get_curr_ctx(dev, MTK_VDEC_CORE);
+
+	if (!mtk_vcodec_is_hw_active(dev)) {
+		mtk_v4l2_vdec_err(ctx, "DEC ISR, VDEC active is not 0x0");
+		return IRQ_HANDLED;
+	}
+
+	dec_done_status = readl(vdec_misc_addr);
+	ctx->irq_status = dec_done_status;
+	if ((dec_done_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) !=
+		MTK_VDEC_IRQ_STATUS_DEC_SUCCESS)
+		return IRQ_HANDLED;
+
+	/* clear interrupt */
+	writel((readl(vdec_misc_addr) | VDEC_IRQ_CFG),
+		dev->reg_base[VDEC_MISC] + VDEC_IRQ_CFG_REG);
+	writel((readl(vdec_misc_addr) & ~VDEC_IRQ_CLR),
+		dev->reg_base[VDEC_MISC] + VDEC_IRQ_CFG_REG);
+
+	wake_up_dec_ctx(ctx, MTK_INST_IRQ_RECEIVED, 0);
+
+	mtk_v4l2_vdec_dbg(3, ctx, "wake up ctx %d, dec_done_status=%x", ctx->id, dec_done_status);
+
+	return IRQ_HANDLED;
+}
+
+static int mtk_vcodec_get_reg_bases(struct mtk_vcodec_dec_dev *dev)
+{
+	struct platform_device *pdev = dev->plat_dev;
+	int reg_num, i;
+	struct resource *res;
+	bool has_vdecsys_reg;
+	int num_max_vdec_regs;
+	static const char * const mtk_dec_reg_names[] = {
+		"misc",
+		"ld",
+		"top",
+		"cm",
+		"ad",
+		"av",
+		"pp",
+		"hwd",
+		"hwq",
+		"hwb",
+		"hwg"
+	};
+
+	/*
+	 * If we have reg-names in devicetree, this means that we're on a new
+	 * register organization, which implies that the VDEC_SYS iospace gets
+	 * R/W through a syscon (regmap).
+	 * Here we try to get the "misc" iostart only to check if we have reg-names
+	 */
+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "misc");
+	if (res)
+		has_vdecsys_reg = false;
+	else
+		has_vdecsys_reg = true;
+
+	num_max_vdec_regs = has_vdecsys_reg ? NUM_MAX_VDEC_REG_BASE :
+					      ARRAY_SIZE(mtk_dec_reg_names);
+
+	/* Sizeof(u32) * 4 bytes for each register base. */
+	reg_num = of_property_count_elems_of_size(pdev->dev.of_node, "reg",
+						  sizeof(u32) * 4);
+	if (reg_num <= 0 || reg_num > num_max_vdec_regs) {
+		dev_err(&pdev->dev, "Invalid register property size: %d\n", reg_num);
+		return -EINVAL;
+	}
+
+	if (has_vdecsys_reg) {
+		for (i = 0; i < reg_num; i++) {
+			dev->reg_base[i] = devm_platform_ioremap_resource(pdev, i);
+			if (IS_ERR(dev->reg_base[i]))
+				return PTR_ERR(dev->reg_base[i]);
+
+			dev_dbg(&pdev->dev, "reg[%d] base=%p", i, dev->reg_base[i]);
+		}
+	} else {
+		for (i = 0; i < reg_num; i++) {
+			dev->reg_base[i+1] = devm_platform_ioremap_resource_byname(pdev, mtk_dec_reg_names[i]);
+			if (IS_ERR(dev->reg_base[i+1]))
+				return PTR_ERR(dev->reg_base[i+1]);
+
+			dev_dbg(&pdev->dev, "reg[%d] base=%p", i + 1, dev->reg_base[i + 1]);
+		}
+
+		dev->vdecsys_regmap = syscon_regmap_lookup_by_phandle(pdev->dev.of_node,
+								      "mediatek,vdecsys");
+		if (IS_ERR(dev->vdecsys_regmap)) {
+			dev_err(&pdev->dev, "Missing mediatek,vdecsys property");
+			return PTR_ERR(dev->vdecsys_regmap);
+		}
+	}
+
+	return 0;
+}
+
+static int mtk_vcodec_init_dec_resources(struct mtk_vcodec_dec_dev *dev)
+{
+	struct platform_device *pdev = dev->plat_dev;
+	int ret;
+
+	ret = mtk_vcodec_get_reg_bases(dev);
+	if (ret)
+		return ret;
+
+	if (dev->vdec_pdata->is_subdev_supported)
+		return 0;
+
+	dev->dec_irq = platform_get_irq(pdev, 0);
+	if (dev->dec_irq < 0)
+		return dev->dec_irq;
+
+	irq_set_status_flags(dev->dec_irq, IRQ_NOAUTOEN);
+	ret = devm_request_irq(&pdev->dev, dev->dec_irq,
+			       mtk_vcodec_dec_irq_handler, 0, pdev->name, dev);
+	if (ret) {
+		dev_err(&pdev->dev, "failed to install dev->dec_irq %d (%d)",
+			dev->dec_irq, ret);
+		return ret;
+	}
+
+	ret = mtk_vcodec_init_dec_clk(pdev, &dev->pm);
+	if (ret < 0) {
+		dev_err(&pdev->dev, "failed to get mt vcodec clock source");
+		return ret;
+	}
+
+	pm_runtime_enable(&pdev->dev);
+	return 0;
+}
+
+static int fops_vcodec_open(struct file *file)
+{
+	struct mtk_vcodec_dec_dev *dev = video_drvdata(file);
+	struct mtk_vcodec_dec_ctx *ctx = NULL;
+	int ret = 0, i, hw_count;
+	struct vb2_queue *src_vq;
+
+	ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
+	if (!ctx)
+		return -ENOMEM;
+
+	mutex_lock(&dev->dev_mutex);
+	ctx->id = dev->id_counter++;
+	v4l2_fh_init(&ctx->fh, video_devdata(file));
+	file->private_data = &ctx->fh;
+	v4l2_fh_add(&ctx->fh);
+	INIT_LIST_HEAD(&ctx->list);
+	ctx->dev = dev;
+	if (ctx->dev->vdec_pdata->is_subdev_supported) {
+		hw_count = mtk_vcodec_get_hw_count(ctx, dev);
+		if (!hw_count || !dev->subdev_prob_done) {
+			ret = -EINVAL;
+			goto err_ctrls_setup;
+		}
+
+		ret = dev->subdev_prob_done(dev);
+		if (ret)
+			goto err_ctrls_setup;
+
+		for (i = 0; i < hw_count; i++)
+			init_waitqueue_head(&ctx->queue[i]);
+	} else {
+		init_waitqueue_head(&ctx->queue[0]);
+	}
+	mutex_init(&ctx->lock);
+
+	ctx->type = MTK_INST_DECODER;
+	ret = dev->vdec_pdata->ctrls_setup(ctx);
+	if (ret) {
+		mtk_v4l2_vdec_err(ctx, "Failed to setup mt vcodec controls");
+		goto err_ctrls_setup;
+	}
+	ctx->m2m_ctx = v4l2_m2m_ctx_init(dev->m2m_dev_dec, ctx,
+		&mtk_vcodec_dec_queue_init);
+	if (IS_ERR((__force void *)ctx->m2m_ctx)) {
+		ret = PTR_ERR((__force void *)ctx->m2m_ctx);
+		mtk_v4l2_vdec_err(ctx, "Failed to v4l2_m2m_ctx_init() (%d)", ret);
+		goto err_m2m_ctx_init;
+	}
+	src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx,
+				V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
+	ctx->empty_flush_buf.vb.vb2_buf.vb2_queue = src_vq;
+	mtk_vcodec_dec_set_default_params(ctx);
+
+	if (v4l2_fh_is_singular(&ctx->fh)) {
+		/*
+		 * Does nothing if firmware was already loaded.
+		 */
+		ret = mtk_vcodec_fw_load_firmware(dev->fw_handler);
+		if (ret < 0) {
+			/*
+			 * Return 0 if downloading firmware successfully,
+			 * otherwise it is failed
+			 */
+			mtk_v4l2_vdec_err(ctx, "failed to load firmware!");
+			goto err_load_fw;
+		}
+
+		dev->dec_capability =
+			mtk_vcodec_fw_get_vdec_capa(dev->fw_handler);
+
+		mtk_v4l2_vdec_dbg(0, ctx, "decoder capability %x", dev->dec_capability);
+	}
+
+	ctx->dev->vdec_pdata->init_vdec_params(ctx);
+
+	list_add(&ctx->list, &dev->ctx_list);
+	mtk_vcodec_dbgfs_create(ctx);
+
+	mutex_unlock(&dev->dev_mutex);
+	mtk_v4l2_vdec_dbg(0, ctx, "%s decoder [%d]", dev_name(&dev->plat_dev->dev), ctx->id);
+	return ret;
+
+	/* Deinit when failure occurred */
+err_load_fw:
+	v4l2_m2m_ctx_release(ctx->m2m_ctx);
+err_m2m_ctx_init:
+	v4l2_ctrl_handler_free(&ctx->ctrl_hdl);
+err_ctrls_setup:
+	v4l2_fh_del(&ctx->fh);
+	v4l2_fh_exit(&ctx->fh);
+	kfree(ctx);
+	mutex_unlock(&dev->dev_mutex);
+
+	return ret;
+}
+
+static int fops_vcodec_release(struct file *file)
+{
+	struct mtk_vcodec_dec_dev *dev = video_drvdata(file);
+	struct mtk_vcodec_dec_ctx *ctx = fh_to_dec_ctx(file->private_data);
+
+	mtk_v4l2_vdec_dbg(0, ctx, "[%d] decoder", ctx->id);
+	mutex_lock(&dev->dev_mutex);
+
+	/*
+	 * Call v4l2_m2m_ctx_release before mtk_vcodec_dec_release. First, it
+	 * makes sure the worker thread is not running after vdec_if_deinit.
+	 * Second, the decoder will be flushed and all the buffers will be
+	 * returned in stop_streaming.
+	 */
+	v4l2_m2m_ctx_release(ctx->m2m_ctx);
+	mtk_vcodec_dec_release(ctx);
+
+	v4l2_fh_del(&ctx->fh);
+	v4l2_fh_exit(&ctx->fh);
+	v4l2_ctrl_handler_free(&ctx->ctrl_hdl);
+
+	mtk_vcodec_dbgfs_remove(dev, ctx->id);
+	list_del_init(&ctx->list);
+	kfree(ctx);
+	mutex_unlock(&dev->dev_mutex);
+	return 0;
+}
+
+static const struct v4l2_file_operations mtk_vcodec_fops = {
+	.owner		= THIS_MODULE,
+	.open		= fops_vcodec_open,
+	.release	= fops_vcodec_release,
+	.poll		= v4l2_m2m_fop_poll,
+	.unlocked_ioctl	= video_ioctl2,
+	.mmap		= v4l2_m2m_fop_mmap,
+};
+
+static int mtk_vcodec_probe(struct platform_device *pdev)
+{
+	struct mtk_vcodec_dec_dev *dev;
+	struct video_device *vfd_dec;
+	phandle rproc_phandle;
+	enum mtk_vcodec_fw_type fw_type;
+	int i, ret;
+
+	dev = devm_kzalloc(&pdev->dev, sizeof(*dev), GFP_KERNEL);
+	if (!dev)
+		return -ENOMEM;
+
+	INIT_LIST_HEAD(&dev->ctx_list);
+	dev->plat_dev = pdev;
+
+	dev->vdec_pdata = of_device_get_match_data(&pdev->dev);
+	if (!of_property_read_u32(pdev->dev.of_node, "mediatek,vpu",
+				  &rproc_phandle)) {
+		fw_type = VPU;
+	} else if (!of_property_read_u32(pdev->dev.of_node, "mediatek,scp",
+					 &rproc_phandle)) {
+		fw_type = SCP;
+	} else {
+		dev_dbg(&pdev->dev, "Could not get vdec IPI device");
+		return -ENODEV;
+	}
+	dma_set_max_seg_size(&pdev->dev, UINT_MAX);
+
+	dev->fw_handler = mtk_vcodec_fw_select(dev, fw_type, DECODER);
+	if (IS_ERR(dev->fw_handler))
+		return PTR_ERR(dev->fw_handler);
+
+	ret = mtk_vcodec_init_dec_resources(dev);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to init dec resources");
+		goto err_dec_pm;
+	}
+
+	if (IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch)) {
+		dev->core_workqueue =
+			alloc_ordered_workqueue("core-decoder",
+						WQ_MEM_RECLAIM | WQ_FREEZABLE);
+		if (!dev->core_workqueue) {
+			dev_dbg(&pdev->dev, "Failed to create core workqueue");
+			ret = -EINVAL;
+			goto err_res;
+		}
+	}
+
+	for (i = 0; i < MTK_VDEC_HW_MAX; i++)
+		mutex_init(&dev->dec_mutex[i]);
+	mutex_init(&dev->dev_mutex);
+	spin_lock_init(&dev->irqlock);
+
+	snprintf(dev->v4l2_dev.name, sizeof(dev->v4l2_dev.name), "%s",
+		"[/MTK_V4L2_VDEC]");
+
+	ret = v4l2_device_register(&pdev->dev, &dev->v4l2_dev);
+	if (ret) {
+		dev_err(&pdev->dev, "v4l2_device_register err=%d", ret);
+		goto err_core_workq;
+	}
+
+	vfd_dec = video_device_alloc();
+	if (!vfd_dec) {
+		dev_err(&pdev->dev, "Failed to allocate video device");
+		ret = -ENOMEM;
+		goto err_dec_alloc;
+	}
+	vfd_dec->fops		= &mtk_vcodec_fops;
+	vfd_dec->ioctl_ops	= &mtk_vdec_ioctl_ops;
+	vfd_dec->release	= video_device_release;
+	vfd_dec->lock		= &dev->dev_mutex;
+	vfd_dec->v4l2_dev	= &dev->v4l2_dev;
+	vfd_dec->vfl_dir	= VFL_DIR_M2M;
+	vfd_dec->device_caps	= V4L2_CAP_VIDEO_M2M_MPLANE |
+			V4L2_CAP_STREAMING;
+
+	snprintf(vfd_dec->name, sizeof(vfd_dec->name), "%s",
+		MTK_VCODEC_DEC_NAME);
+	video_set_drvdata(vfd_dec, dev);
+	dev->vfd_dec = vfd_dec;
+	platform_set_drvdata(pdev, dev);
+
+	dev->m2m_dev_dec = v4l2_m2m_init(&mtk_vdec_m2m_ops);
+	if (IS_ERR((__force void *)dev->m2m_dev_dec)) {
+		dev_err(&pdev->dev, "Failed to init mem2mem dec device");
+		ret = PTR_ERR((__force void *)dev->m2m_dev_dec);
+		goto err_dec_alloc;
+	}
+
+	dev->decode_workqueue =
+		alloc_ordered_workqueue(MTK_VCODEC_DEC_NAME,
+			WQ_MEM_RECLAIM | WQ_FREEZABLE);
+	if (!dev->decode_workqueue) {
+		dev_err(&pdev->dev, "Failed to create decode workqueue");
+		ret = -EINVAL;
+		goto err_event_workq;
+	}
+
+	if (dev->vdec_pdata->is_subdev_supported) {
+		ret = of_platform_populate(pdev->dev.of_node, NULL, NULL,
+					   &pdev->dev);
+		if (ret) {
+			dev_err(&pdev->dev, "Main device of_platform_populate failed.");
+			goto err_reg_cont;
+		}
+	} else {
+		set_bit(MTK_VDEC_CORE, dev->subdev_bitmap);
+	}
+
+	atomic_set(&dev->dec_active_cnt, 0);
+	memset(dev->vdec_racing_info, 0, sizeof(dev->vdec_racing_info));
+	mutex_init(&dev->dec_racing_info_mutex);
+
+	ret = video_register_device(vfd_dec, VFL_TYPE_VIDEO, -1);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to register video device");
+		goto err_reg_cont;
+	}
+
+	if (dev->vdec_pdata->uses_stateless_api) {
+		v4l2_disable_ioctl(vfd_dec, VIDIOC_DECODER_CMD);
+		v4l2_disable_ioctl(vfd_dec, VIDIOC_TRY_DECODER_CMD);
+
+		dev->mdev_dec.dev = &pdev->dev;
+		strscpy(dev->mdev_dec.model, MTK_VCODEC_DEC_NAME,
+			sizeof(dev->mdev_dec.model));
+
+		media_device_init(&dev->mdev_dec);
+		dev->mdev_dec.ops = &mtk_vcodec_media_ops;
+		dev->v4l2_dev.mdev = &dev->mdev_dec;
+
+		ret = v4l2_m2m_register_media_controller(dev->m2m_dev_dec, dev->vfd_dec,
+							 MEDIA_ENT_F_PROC_VIDEO_DECODER);
+		if (ret) {
+			dev_err(&pdev->dev, "Failed to register media controller");
+			goto err_dec_mem_init;
+		}
+
+		ret = media_device_register(&dev->mdev_dec);
+		if (ret) {
+			dev_err(&pdev->dev, "Failed to register media device");
+			goto err_media_reg;
+		}
+
+		dev_dbg(&pdev->dev, "media registered as /dev/media%d", vfd_dec->minor);
+	}
+
+	mtk_vcodec_dbgfs_init(dev, false);
+	dev_dbg(&pdev->dev, "decoder registered as /dev/video%d", vfd_dec->minor);
+
+	return 0;
+
+err_media_reg:
+	v4l2_m2m_unregister_media_controller(dev->m2m_dev_dec);
+err_dec_mem_init:
+	video_unregister_device(vfd_dec);
+err_reg_cont:
+	if (dev->vdec_pdata->uses_stateless_api)
+		media_device_cleanup(&dev->mdev_dec);
+	destroy_workqueue(dev->decode_workqueue);
+err_event_workq:
+	v4l2_m2m_release(dev->m2m_dev_dec);
+err_dec_alloc:
+	v4l2_device_unregister(&dev->v4l2_dev);
+err_core_workq:
+	if (IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch))
+		destroy_workqueue(dev->core_workqueue);
+err_res:
+	if (!dev->vdec_pdata->is_subdev_supported)
+		pm_runtime_disable(dev->pm.dev);
+err_dec_pm:
+	mtk_vcodec_fw_release(dev->fw_handler);
+	return ret;
+}
+
+static const struct of_device_id mtk_vcodec_match[] = {
+	{
+		.compatible = "mediatek,mt8173-vcodec-dec",
+		.data = &mtk_vdec_8173_pdata,
+	},
+	{
+		.compatible = "mediatek,mt8183-vcodec-dec",
+		.data = &mtk_vdec_8183_pdata,
+	},
+	{
+		.compatible = "mediatek,mt8192-vcodec-dec",
+		.data = &mtk_lat_sig_core_pdata,
+	},
+	{
+		.compatible = "mediatek,mt8186-vcodec-dec",
+		.data = &mtk_vdec_single_core_pdata,
+	},
+	{
+		.compatible = "mediatek,mt8195-vcodec-dec",
+		.data = &mtk_lat_sig_core_pdata,
+	},
+	{
+		.compatible = "mediatek,mt8188-vcodec-dec",
+		.data = &mtk_lat_sig_core_pdata,
+	},
+	{},
+};
+
+MODULE_DEVICE_TABLE(of, mtk_vcodec_match);
+
+static void mtk_vcodec_dec_remove(struct platform_device *pdev)
+{
+	struct mtk_vcodec_dec_dev *dev = platform_get_drvdata(pdev);
+
+	destroy_workqueue(dev->decode_workqueue);
+
+	if (media_devnode_is_registered(dev->mdev_dec.devnode)) {
+		media_device_unregister(&dev->mdev_dec);
+		v4l2_m2m_unregister_media_controller(dev->m2m_dev_dec);
+		media_device_cleanup(&dev->mdev_dec);
+	}
+
+	if (dev->m2m_dev_dec)
+		v4l2_m2m_release(dev->m2m_dev_dec);
+
+	if (dev->vfd_dec)
+		video_unregister_device(dev->vfd_dec);
+
+	mtk_vcodec_dbgfs_deinit(&dev->dbgfs);
+	v4l2_device_unregister(&dev->v4l2_dev);
+	if (!dev->vdec_pdata->is_subdev_supported)
+		pm_runtime_disable(dev->pm.dev);
+	mtk_vcodec_fw_release(dev->fw_handler);
+}
+
+static struct platform_driver mtk_vcodec_dec_driver = {
+	.probe	= mtk_vcodec_probe,
+	.remove_new = mtk_vcodec_dec_remove,
+	.driver	= {
+		.name	= MTK_VCODEC_DEC_NAME,
+		.of_match_table = mtk_vcodec_match,
+	},
+};
+
+module_platform_driver(mtk_vcodec_dec_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Mediatek video codec V4L2 decoder driver");
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h
new file mode 100644
index 0000000000..7e36b2c69b
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h
@@ -0,0 +1,324 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2023 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#ifndef _MTK_VCODEC_DEC_DRV_H_
+#define _MTK_VCODEC_DEC_DRV_H_
+
+#include "../common/mtk_vcodec_cmn_drv.h"
+#include "../common/mtk_vcodec_dbgfs.h"
+#include "../common/mtk_vcodec_fw_priv.h"
+#include "../common/mtk_vcodec_util.h"
+#include "vdec_msg_queue.h"
+
+#define MTK_VCODEC_DEC_NAME	"mtk-vcodec-dec"
+
+#define IS_VDEC_LAT_ARCH(hw_arch) ((hw_arch) >= MTK_VDEC_LAT_SINGLE_CORE)
+#define IS_VDEC_INNER_RACING(capability) ((capability) & MTK_VCODEC_INNER_RACING)
+
+/*
+ * enum mtk_vdec_format_types - Structure used to get supported
+ *		  format types according to decoder capability
+ */
+enum mtk_vdec_format_types {
+	MTK_VDEC_FORMAT_MM21 = 0x20,
+	MTK_VDEC_FORMAT_MT21C = 0x40,
+	MTK_VDEC_FORMAT_H264_SLICE = 0x100,
+	MTK_VDEC_FORMAT_VP8_FRAME = 0x200,
+	MTK_VDEC_FORMAT_VP9_FRAME = 0x400,
+	MTK_VDEC_FORMAT_AV1_FRAME = 0x800,
+	MTK_VDEC_FORMAT_HEVC_FRAME = 0x1000,
+	MTK_VCODEC_INNER_RACING = 0x20000,
+	MTK_VDEC_IS_SUPPORT_10BIT = 0x40000,
+};
+
+/*
+ * enum mtk_vdec_hw_count - Supported hardware count
+ */
+enum mtk_vdec_hw_count {
+	MTK_VDEC_NO_HW = 0,
+	MTK_VDEC_ONE_CORE,
+	MTK_VDEC_ONE_LAT_ONE_CORE,
+	MTK_VDEC_MAX_HW_COUNT,
+};
+
+/*
+ * enum mtk_vdec_hw_arch - Used to separate different hardware architecture
+ */
+enum mtk_vdec_hw_arch {
+	MTK_VDEC_PURE_SINGLE_CORE,
+	MTK_VDEC_LAT_SINGLE_CORE,
+};
+
+/**
+ * struct vdec_pic_info  - picture size information
+ * @pic_w: picture width
+ * @pic_h: picture height
+ * @buf_w: picture buffer width (64 aligned up from pic_w)
+ * @buf_h: picture buffer heiht (64 aligned up from pic_h)
+ * @fb_sz: bitstream size of each plane
+ * E.g. suppose picture size is 176x144,
+ *      buffer size will be aligned to 176x160.
+ * @cap_fourcc: fourcc number(may changed when resolution change)
+ * @reserved: align struct to 64-bit in order to adjust 32-bit and 64-bit os.
+ */
+struct vdec_pic_info {
+	unsigned int pic_w;
+	unsigned int pic_h;
+	unsigned int buf_w;
+	unsigned int buf_h;
+	unsigned int fb_sz[VIDEO_MAX_PLANES];
+	unsigned int cap_fourcc;
+	unsigned int reserved;
+};
+
+/**
+ * struct mtk_vcodec_dec_pdata - compatible data for each IC
+ * @init_vdec_params: init vdec params
+ * @ctrls_setup: init vcodec dec ctrls
+ * @worker: worker to start a decode job
+ * @flush_decoder: function that flushes the decoder
+ * @get_cap_buffer: get capture buffer from capture queue
+ * @cap_to_disp: put capture buffer to disp list for lat and core arch
+ * @vdec_vb2_ops: struct vb2_ops
+ *
+ * @vdec_formats: supported video decoder formats
+ * @num_formats: count of video decoder formats
+ * @default_out_fmt: default output buffer format
+ * @default_cap_fmt: default capture buffer format
+ *
+ * @hw_arch: hardware arch is used to separate pure_sin_core and lat_sin_core
+ *
+ * @is_subdev_supported: whether support parent-node architecture(subdev)
+ * @uses_stateless_api: whether the decoder uses the stateless API with requests
+ */
+struct mtk_vcodec_dec_pdata {
+	void (*init_vdec_params)(struct mtk_vcodec_dec_ctx *ctx);
+	int (*ctrls_setup)(struct mtk_vcodec_dec_ctx *ctx);
+	void (*worker)(struct work_struct *work);
+	int (*flush_decoder)(struct mtk_vcodec_dec_ctx *ctx);
+	struct vdec_fb *(*get_cap_buffer)(struct mtk_vcodec_dec_ctx *ctx);
+	void (*cap_to_disp)(struct mtk_vcodec_dec_ctx *ctx, int error,
+			    struct media_request *src_buf_req);
+
+	const struct vb2_ops *vdec_vb2_ops;
+
+	const struct mtk_video_fmt *vdec_formats;
+	const int *num_formats;
+	const struct mtk_video_fmt *default_out_fmt;
+	const struct mtk_video_fmt *default_cap_fmt;
+
+	enum mtk_vdec_hw_arch hw_arch;
+
+	bool is_subdev_supported;
+	bool uses_stateless_api;
+};
+
+/**
+ * struct mtk_vcodec_dec_ctx - Context (instance) private data.
+ *
+ * @type: type of decoder instance
+ * @dev: pointer to the mtk_vcodec_dec_dev of the device
+ * @list: link to ctx_list of mtk_vcodec_dec_dev
+ *
+ * @fh: struct v4l2_fh
+ * @m2m_ctx: pointer to the v4l2_m2m_ctx of the context
+ * @q_data: store information of input and output queue of the context
+ * @id: index of the context that this structure describes
+ * @state: state of the context
+ *
+ * @dec_if: hooked decoder driver interface
+ * @drv_handle: driver handle for specific decode/encode instance
+ *
+ * @picinfo: store picture info after header parsing
+ * @dpb_size: store dpb count after header parsing
+ *
+ * @int_cond: variable used by the waitqueue
+ * @int_type: type of the last interrupt
+ * @queue: waitqueue that can be used to wait for this context to finish
+ * @irq_status: irq status
+ *
+ * @ctrl_hdl: handler for v4l2 framework
+ * @decode_work: worker for the decoding
+ * @last_decoded_picinfo: pic information get from latest decode
+ * @empty_flush_buf: a fake size-0 capture buffer that indicates flush. Used
+ *		     for stateful decoder.
+ * @is_flushing: set to true if flushing is in progress.
+ *
+ * @current_codec: current set input codec, in V4L2 pixel format
+ * @capture_fourcc: capture queue type in V4L2 pixel format
+ *
+ * @colorspace: enum v4l2_colorspace; supplemental to pixelformat
+ * @ycbcr_enc: enum v4l2_ycbcr_encoding, Y'CbCr encoding
+ * @quantization: enum v4l2_quantization, colorspace quantization
+ * @xfer_func: enum v4l2_xfer_func, colorspace transfer function
+ *
+ * @decoded_frame_cnt: number of decoded frames
+ * @lock: protect variables accessed by V4L2 threads and worker thread such as
+ *	  mtk_video_dec_buf.
+ * @hw_id: hardware index used to identify different hardware.
+ *
+ * @msg_queue: msg queue used to store lat buffer information.
+ * @vpu_inst: vpu instance pointer.
+ *
+ * @is_10bit_bitstream: set to true if it's 10bit bitstream
+ */
+struct mtk_vcodec_dec_ctx {
+	enum mtk_instance_type type;
+	struct mtk_vcodec_dec_dev *dev;
+	struct list_head list;
+
+	struct v4l2_fh fh;
+	struct v4l2_m2m_ctx *m2m_ctx;
+	struct mtk_q_data q_data[2];
+	int id;
+	enum mtk_instance_state state;
+
+	const struct vdec_common_if *dec_if;
+	void *drv_handle;
+
+	struct vdec_pic_info picinfo;
+	int dpb_size;
+
+	int int_cond[MTK_VDEC_HW_MAX];
+	int int_type[MTK_VDEC_HW_MAX];
+	wait_queue_head_t queue[MTK_VDEC_HW_MAX];
+	unsigned int irq_status;
+
+	struct v4l2_ctrl_handler ctrl_hdl;
+	struct work_struct decode_work;
+	struct vdec_pic_info last_decoded_picinfo;
+	struct v4l2_m2m_buffer empty_flush_buf;
+	bool is_flushing;
+
+	u32 current_codec;
+	u32 capture_fourcc;
+
+	enum v4l2_colorspace colorspace;
+	enum v4l2_ycbcr_encoding ycbcr_enc;
+	enum v4l2_quantization quantization;
+	enum v4l2_xfer_func xfer_func;
+
+	int decoded_frame_cnt;
+	struct mutex lock;
+	int hw_id;
+
+	struct vdec_msg_queue msg_queue;
+	void *vpu_inst;
+
+	bool is_10bit_bitstream;
+};
+
+/**
+ * struct mtk_vcodec_dec_dev - driver data
+ * @v4l2_dev: V4L2 device to register video devices for.
+ * @vfd_dec: Video device for decoder
+ * @mdev_dec: Media device for decoder
+ *
+ * @m2m_dev_dec: m2m device for decoder
+ * @plat_dev: platform device
+ * @ctx_list: list of struct mtk_vcodec_ctx
+ * @curr_ctx: The context that is waiting for codec hardware
+ *
+ * @reg_base: Mapped address of MTK Vcodec registers.
+ * @vdec_pdata: decoder IC-specific data
+ * @vdecsys_regmap: VDEC_SYS register space passed through syscon
+ *
+ * @fw_handler: used to communicate with the firmware.
+ * @id_counter: used to identify current opened instance
+ *
+ * @dec_mutex: decoder hardware lock
+ * @dev_mutex: video_device lock
+ * @decode_workqueue: decode work queue
+ *
+ * @irqlock: protect data access by irq handler and work thread
+ * @dec_irq: decoder irq resource
+ *
+ * @pm: power management control
+ * @dec_capability: used to identify decode capability, ex: 4k
+ *
+ * @core_workqueue: queue used for core hardware decode
+ *
+ * @subdev_dev: subdev hardware device
+ * @subdev_prob_done: check whether all used hw device is prob done
+ * @subdev_bitmap: used to record hardware is ready or not
+ *
+ * @dec_active_cnt: used to mark whether need to record register value
+ * @vdec_racing_info: record register value
+ * @dec_racing_info_mutex: mutex lock used for inner racing mode
+ * @dbgfs: debug log related information
+ */
+struct mtk_vcodec_dec_dev {
+	struct v4l2_device v4l2_dev;
+	struct video_device *vfd_dec;
+	struct media_device mdev_dec;
+
+	struct v4l2_m2m_dev *m2m_dev_dec;
+	struct platform_device *plat_dev;
+	struct list_head ctx_list;
+	struct mtk_vcodec_dec_ctx *curr_ctx;
+
+	void __iomem *reg_base[NUM_MAX_VCODEC_REG_BASE];
+	const struct mtk_vcodec_dec_pdata *vdec_pdata;
+	struct regmap *vdecsys_regmap;
+
+	struct mtk_vcodec_fw *fw_handler;
+	u64 id_counter;
+
+	/* decoder hardware mutex lock */
+	struct mutex dec_mutex[MTK_VDEC_HW_MAX];
+	struct mutex dev_mutex;
+	struct workqueue_struct *decode_workqueue;
+
+	spinlock_t irqlock;
+	int dec_irq;
+
+	struct mtk_vcodec_pm pm;
+	unsigned int dec_capability;
+
+	struct workqueue_struct *core_workqueue;
+
+	void *subdev_dev[MTK_VDEC_HW_MAX];
+	int (*subdev_prob_done)(struct mtk_vcodec_dec_dev *vdec_dev);
+	DECLARE_BITMAP(subdev_bitmap, MTK_VDEC_HW_MAX);
+
+	atomic_t dec_active_cnt;
+	u32 vdec_racing_info[132];
+	/* Protects access to vdec_racing_info data */
+	struct mutex dec_racing_info_mutex;
+	struct mtk_vcodec_dbgfs dbgfs;
+};
+
+static inline struct mtk_vcodec_dec_ctx *fh_to_dec_ctx(struct v4l2_fh *fh)
+{
+	return container_of(fh, struct mtk_vcodec_dec_ctx, fh);
+}
+
+static inline struct mtk_vcodec_dec_ctx *ctrl_to_dec_ctx(struct v4l2_ctrl *ctrl)
+{
+	return container_of(ctrl->handler, struct mtk_vcodec_dec_ctx, ctrl_hdl);
+}
+
+/* Wake up context wait_queue */
+static inline void
+wake_up_dec_ctx(struct mtk_vcodec_dec_ctx *ctx, unsigned int reason, unsigned int hw_id)
+{
+	ctx->int_cond[hw_id] = 1;
+	ctx->int_type[hw_id] = reason;
+	wake_up_interruptible(&ctx->queue[hw_id]);
+}
+
+#define mtk_vdec_err(ctx, fmt, args...)                               \
+	mtk_vcodec_err((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args)
+
+#define mtk_vdec_debug(ctx, fmt, args...)                             \
+	mtk_vcodec_debug((ctx)->id, (ctx)->dev->plat_dev, fmt, ##args)
+
+#define mtk_v4l2_vdec_err(ctx, fmt, args...) mtk_v4l2_err((ctx)->dev->plat_dev, fmt, ##args)
+
+#define mtk_v4l2_vdec_dbg(level, ctx, fmt, args...)             \
+	mtk_v4l2_debug((ctx)->dev->plat_dev, level, fmt, ##args)
+
+#endif /* _MTK_VCODEC_DEC_DRV_H_ */
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c
new file mode 100644
index 0000000000..881d5de41e
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.c
@@ -0,0 +1,201 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include <linux/interrupt.h>
+#include <linux/irq.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/slab.h>
+
+#include "mtk_vcodec_dec.h"
+#include "mtk_vcodec_dec_hw.h"
+#include "mtk_vcodec_dec_pm.h"
+#include "../common/mtk_vcodec_intr.h"
+
+static const struct of_device_id mtk_vdec_hw_match[] = {
+	{
+		.compatible = "mediatek,mtk-vcodec-lat",
+		.data = (void *)MTK_VDEC_LAT0,
+	},
+	{
+		.compatible = "mediatek,mtk-vcodec-core",
+		.data = (void *)MTK_VDEC_CORE,
+	},
+	{
+		.compatible = "mediatek,mtk-vcodec-lat-soc",
+		.data = (void *)MTK_VDEC_LAT_SOC,
+	},
+	{},
+};
+MODULE_DEVICE_TABLE(of, mtk_vdec_hw_match);
+
+static int mtk_vdec_hw_prob_done(struct mtk_vcodec_dec_dev *vdec_dev)
+{
+	struct platform_device *pdev = vdec_dev->plat_dev;
+	struct device_node *subdev_node;
+	enum mtk_vdec_hw_id hw_idx;
+	const struct of_device_id *of_id;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(mtk_vdec_hw_match); i++) {
+		of_id = &mtk_vdec_hw_match[i];
+		subdev_node = of_find_compatible_node(NULL, NULL,
+						      of_id->compatible);
+		if (!subdev_node)
+			continue;
+
+		of_node_put(subdev_node);
+
+		hw_idx = (enum mtk_vdec_hw_id)(uintptr_t)of_id->data;
+		if (!test_bit(hw_idx, vdec_dev->subdev_bitmap)) {
+			dev_err(&pdev->dev, "vdec %d is not ready", hw_idx);
+			return -EAGAIN;
+		}
+	}
+
+	return 0;
+}
+
+static irqreturn_t mtk_vdec_hw_irq_handler(int irq, void *priv)
+{
+	struct mtk_vdec_hw_dev *dev = priv;
+	struct mtk_vcodec_dec_ctx *ctx;
+	u32 cg_status;
+	unsigned int dec_done_status;
+	void __iomem *vdec_misc_addr = dev->reg_base[VDEC_HW_MISC] +
+					VDEC_IRQ_CFG_REG;
+
+	ctx = mtk_vcodec_get_curr_ctx(dev->main_dev, dev->hw_idx);
+
+	/* check if HW active or not */
+	cg_status = readl(dev->reg_base[VDEC_HW_SYS] + VDEC_HW_ACTIVE_ADDR);
+	if (cg_status & VDEC_HW_ACTIVE_MASK) {
+		mtk_v4l2_vdec_err(ctx, "vdec active is not 0x0 (0x%08x)", cg_status);
+		return IRQ_HANDLED;
+	}
+
+	dec_done_status = readl(vdec_misc_addr);
+	if ((dec_done_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS) !=
+	    MTK_VDEC_IRQ_STATUS_DEC_SUCCESS)
+		return IRQ_HANDLED;
+
+	/* clear interrupt */
+	writel(dec_done_status | VDEC_IRQ_CFG, vdec_misc_addr);
+	writel(dec_done_status & ~VDEC_IRQ_CLR, vdec_misc_addr);
+
+	wake_up_dec_ctx(ctx, MTK_INST_IRQ_RECEIVED, dev->hw_idx);
+
+	mtk_v4l2_vdec_dbg(3, ctx, "wake up ctx %d, dec_done_status=%x",
+			  ctx->id, dec_done_status);
+
+	return IRQ_HANDLED;
+}
+
+static int mtk_vdec_hw_init_irq(struct mtk_vdec_hw_dev *dev)
+{
+	struct platform_device *pdev = dev->plat_dev;
+	int ret;
+
+	dev->dec_irq = platform_get_irq(pdev, 0);
+	if (dev->dec_irq < 0)
+		return dev->dec_irq;
+
+	irq_set_status_flags(dev->dec_irq, IRQ_NOAUTOEN);
+	ret = devm_request_irq(&pdev->dev, dev->dec_irq,
+			       mtk_vdec_hw_irq_handler, 0, pdev->name, dev);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to install dev->dec_irq %d (%d)",
+			dev->dec_irq, ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int mtk_vdec_hw_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct mtk_vdec_hw_dev *subdev_dev;
+	struct mtk_vcodec_dec_dev *main_dev;
+	const struct of_device_id *of_id;
+	int hw_idx;
+	int ret;
+
+	if (!dev->parent) {
+		dev_err(dev, "no parent for hardware devices.\n");
+		return -ENODEV;
+	}
+
+	main_dev = dev_get_drvdata(dev->parent);
+	if (!main_dev) {
+		dev_err(dev, "failed to get parent driver data");
+		return -EINVAL;
+	}
+
+	subdev_dev = devm_kzalloc(dev, sizeof(*subdev_dev), GFP_KERNEL);
+	if (!subdev_dev)
+		return -ENOMEM;
+
+	subdev_dev->plat_dev = pdev;
+	ret = mtk_vcodec_init_dec_clk(pdev, &subdev_dev->pm);
+	if (ret)
+		return ret;
+
+	ret = devm_pm_runtime_enable(&pdev->dev);
+	if (ret)
+		return ret;
+
+	of_id = of_match_device(mtk_vdec_hw_match, dev);
+	if (!of_id) {
+		dev_err(dev, "Can't get vdec subdev id.\n");
+		return -EINVAL;
+	}
+
+	hw_idx = (enum mtk_vdec_hw_id)(uintptr_t)of_id->data;
+	if (hw_idx >= MTK_VDEC_HW_MAX) {
+		dev_err(dev, "Hardware index %d not correct.\n", hw_idx);
+		return -EINVAL;
+	}
+
+	main_dev->subdev_dev[hw_idx] = subdev_dev;
+	subdev_dev->hw_idx = hw_idx;
+	subdev_dev->main_dev = main_dev;
+	subdev_dev->reg_base[VDEC_HW_SYS] = main_dev->reg_base[VDEC_HW_SYS];
+	set_bit(subdev_dev->hw_idx, main_dev->subdev_bitmap);
+
+	if (IS_SUPPORT_VDEC_HW_IRQ(hw_idx)) {
+		ret = mtk_vdec_hw_init_irq(subdev_dev);
+		if (ret)
+			return ret;
+	}
+
+	subdev_dev->reg_base[VDEC_HW_MISC] =
+		devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR((__force void *)subdev_dev->reg_base[VDEC_HW_MISC])) {
+		ret = PTR_ERR((__force void *)subdev_dev->reg_base[VDEC_HW_MISC]);
+		return ret;
+	}
+
+	if (!main_dev->subdev_prob_done)
+		main_dev->subdev_prob_done = mtk_vdec_hw_prob_done;
+
+	platform_set_drvdata(pdev, subdev_dev);
+	return 0;
+}
+
+static struct platform_driver mtk_vdec_driver = {
+	.probe	= mtk_vdec_hw_probe,
+	.driver	= {
+		.name	= "mtk-vdec-comp",
+		.of_match_table = mtk_vdec_hw_match,
+	},
+};
+module_platform_driver(mtk_vdec_driver);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Mediatek video decoder hardware driver");
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h
new file mode 100644
index 0000000000..83fe8b9428
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_hw.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#ifndef _MTK_VCODEC_DEC_HW_H_
+#define _MTK_VCODEC_DEC_HW_H_
+
+#include <linux/io.h>
+#include <linux/platform_device.h>
+
+#include "mtk_vcodec_dec_drv.h"
+
+#define VDEC_HW_ACTIVE_ADDR 0x0
+#define VDEC_HW_ACTIVE_MASK BIT(4)
+#define VDEC_IRQ_CFG 0x11
+#define VDEC_IRQ_CLR 0x10
+#define VDEC_IRQ_CFG_REG 0xa4
+
+#define IS_SUPPORT_VDEC_HW_IRQ(hw_idx) ((hw_idx) != MTK_VDEC_LAT_SOC)
+
+/**
+ * enum mtk_vdec_hw_reg_idx - subdev hardware register base index
+ * @VDEC_HW_SYS : vdec soc register index
+ * @VDEC_HW_MISC: vdec misc register index
+ * @VDEC_HW_MAX : vdec supported max register index
+ */
+enum mtk_vdec_hw_reg_idx {
+	VDEC_HW_SYS,
+	VDEC_HW_MISC,
+	VDEC_HW_MAX
+};
+
+/**
+ * struct mtk_vdec_hw_dev - vdec hardware driver data
+ * @plat_dev: platform device
+ * @main_dev: main device
+ * @reg_base: mapped address of MTK Vcodec registers.
+ *
+ * @curr_ctx: the context that is waiting for codec hardware
+ *
+ * @dec_irq : decoder irq resource
+ * @pm      : power management control
+ * @hw_idx  : each hardware index
+ */
+struct mtk_vdec_hw_dev {
+	struct platform_device *plat_dev;
+	struct mtk_vcodec_dec_dev *main_dev;
+	void __iomem *reg_base[VDEC_HW_MAX];
+
+	struct mtk_vcodec_dec_ctx *curr_ctx;
+
+	int dec_irq;
+	struct mtk_vcodec_pm pm;
+	int hw_idx;
+};
+
+#endif /* _MTK_VCODEC_DEC_HW_H_ */
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c
new file mode 100644
index 0000000000..aefd3e9e30
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.c
@@ -0,0 +1,262 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#include <linux/clk.h>
+#include <linux/interrupt.h>
+#include <linux/of.h>
+#include <linux/pm_runtime.h>
+
+#include "mtk_vcodec_dec_hw.h"
+#include "mtk_vcodec_dec_pm.h"
+
+int mtk_vcodec_init_dec_clk(struct platform_device *pdev, struct mtk_vcodec_pm *pm)
+{
+	struct mtk_vcodec_clk *dec_clk;
+	struct mtk_vcodec_clk_info *clk_info;
+	int i = 0, ret;
+
+	dec_clk = &pm->vdec_clk;
+	pm->dev = &pdev->dev;
+
+	dec_clk->clk_num =
+		of_property_count_strings(pdev->dev.of_node, "clock-names");
+	if (dec_clk->clk_num > 0) {
+		dec_clk->clk_info = devm_kcalloc(&pdev->dev,
+			dec_clk->clk_num, sizeof(*clk_info),
+			GFP_KERNEL);
+		if (!dec_clk->clk_info)
+			return -ENOMEM;
+	} else {
+		dev_err(&pdev->dev, "Failed to get vdec clock count");
+		return -EINVAL;
+	}
+
+	for (i = 0; i < dec_clk->clk_num; i++) {
+		clk_info = &dec_clk->clk_info[i];
+		ret = of_property_read_string_index(pdev->dev.of_node,
+			"clock-names", i, &clk_info->clk_name);
+		if (ret) {
+			dev_err(&pdev->dev, "Failed to get clock name id = %d", i);
+			return ret;
+		}
+		clk_info->vcodec_clk = devm_clk_get(&pdev->dev,
+			clk_info->clk_name);
+		if (IS_ERR(clk_info->vcodec_clk)) {
+			dev_err(&pdev->dev, "devm_clk_get (%d)%s fail", i, clk_info->clk_name);
+			return PTR_ERR(clk_info->vcodec_clk);
+		}
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(mtk_vcodec_init_dec_clk);
+
+static int mtk_vcodec_dec_pw_on(struct mtk_vcodec_pm *pm)
+{
+	int ret;
+
+	ret = pm_runtime_resume_and_get(pm->dev);
+	if (ret)
+		dev_err(pm->dev, "pm_runtime_resume_and_get fail %d", ret);
+
+	return ret;
+}
+
+static void mtk_vcodec_dec_pw_off(struct mtk_vcodec_pm *pm)
+{
+	int ret;
+
+	ret = pm_runtime_put(pm->dev);
+	if (ret && ret != -EAGAIN)
+		dev_err(pm->dev, "pm_runtime_put fail %d", ret);
+}
+
+static void mtk_vcodec_dec_clock_on(struct mtk_vcodec_pm *pm)
+{
+	struct mtk_vcodec_clk *dec_clk;
+	int ret, i;
+
+	dec_clk = &pm->vdec_clk;
+	for (i = 0; i < dec_clk->clk_num; i++) {
+		ret = clk_prepare_enable(dec_clk->clk_info[i].vcodec_clk);
+		if (ret) {
+			dev_err(pm->dev, "clk_prepare_enable %d %s fail %d", i,
+				dec_clk->clk_info[i].clk_name, ret);
+			goto error;
+		}
+	}
+
+	return;
+error:
+	for (i -= 1; i >= 0; i--)
+		clk_disable_unprepare(dec_clk->clk_info[i].vcodec_clk);
+}
+
+static void mtk_vcodec_dec_clock_off(struct mtk_vcodec_pm *pm)
+{
+	struct mtk_vcodec_clk *dec_clk;
+	int i;
+
+	dec_clk = &pm->vdec_clk;
+	for (i = dec_clk->clk_num - 1; i >= 0; i--)
+		clk_disable_unprepare(dec_clk->clk_info[i].vcodec_clk);
+}
+
+static void mtk_vcodec_dec_enable_irq(struct mtk_vcodec_dec_dev *vdec_dev, int hw_idx)
+{
+	struct mtk_vdec_hw_dev *subdev_dev;
+
+	if (!test_bit(hw_idx, vdec_dev->subdev_bitmap))
+		return;
+
+	if (vdec_dev->vdec_pdata->is_subdev_supported) {
+		subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx);
+		if (subdev_dev)
+			enable_irq(subdev_dev->dec_irq);
+		else
+			dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n");
+	} else {
+		enable_irq(vdec_dev->dec_irq);
+	}
+}
+
+static void mtk_vcodec_dec_disable_irq(struct mtk_vcodec_dec_dev *vdec_dev, int hw_idx)
+{
+	struct mtk_vdec_hw_dev *subdev_dev;
+
+	if (!test_bit(hw_idx, vdec_dev->subdev_bitmap))
+		return;
+
+	if (vdec_dev->vdec_pdata->is_subdev_supported) {
+		subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx);
+		if (subdev_dev)
+			disable_irq(subdev_dev->dec_irq);
+		else
+			dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n");
+	} else {
+		disable_irq(vdec_dev->dec_irq);
+	}
+}
+
+static void mtk_vcodec_load_racing_info(struct mtk_vcodec_dec_ctx *ctx)
+{
+	void __iomem *vdec_racing_addr;
+	int j;
+
+	mutex_lock(&ctx->dev->dec_racing_info_mutex);
+	if (atomic_inc_return(&ctx->dev->dec_active_cnt) == 1) {
+		vdec_racing_addr = ctx->dev->reg_base[VDEC_MISC] + 0x100;
+		for (j = 0; j < 132; j++)
+			writel(ctx->dev->vdec_racing_info[j], vdec_racing_addr + j * 4);
+	}
+	mutex_unlock(&ctx->dev->dec_racing_info_mutex);
+}
+
+static void mtk_vcodec_record_racing_info(struct mtk_vcodec_dec_ctx *ctx)
+{
+	void __iomem *vdec_racing_addr;
+	int j;
+
+	mutex_lock(&ctx->dev->dec_racing_info_mutex);
+	if (atomic_dec_and_test(&ctx->dev->dec_active_cnt)) {
+		vdec_racing_addr = ctx->dev->reg_base[VDEC_MISC] + 0x100;
+		for (j = 0; j < 132; j++)
+			ctx->dev->vdec_racing_info[j] = readl(vdec_racing_addr + j * 4);
+	}
+	mutex_unlock(&ctx->dev->dec_racing_info_mutex);
+}
+
+static struct mtk_vcodec_pm *mtk_vcodec_dec_get_pm(struct mtk_vcodec_dec_dev *vdec_dev,
+						   int hw_idx)
+{
+	struct mtk_vdec_hw_dev *subdev_dev;
+
+	if (!test_bit(hw_idx, vdec_dev->subdev_bitmap))
+		return NULL;
+
+	if (vdec_dev->vdec_pdata->is_subdev_supported) {
+		subdev_dev = mtk_vcodec_get_hw_dev(vdec_dev, hw_idx);
+		if (subdev_dev)
+			return &subdev_dev->pm;
+
+		dev_err(&vdec_dev->plat_dev->dev, "Failed to get hw dev\n");
+		return NULL;
+	}
+
+	return &vdec_dev->pm;
+}
+
+static void mtk_vcodec_dec_child_dev_on(struct mtk_vcodec_dec_dev *vdec_dev,
+					int hw_idx)
+{
+	struct mtk_vcodec_pm *pm;
+
+	pm = mtk_vcodec_dec_get_pm(vdec_dev, hw_idx);
+	if (pm) {
+		mtk_vcodec_dec_pw_on(pm);
+		mtk_vcodec_dec_clock_on(pm);
+	}
+
+	if (hw_idx == MTK_VDEC_LAT0) {
+		pm = mtk_vcodec_dec_get_pm(vdec_dev, MTK_VDEC_LAT_SOC);
+		if (pm) {
+			mtk_vcodec_dec_pw_on(pm);
+			mtk_vcodec_dec_clock_on(pm);
+		}
+	}
+}
+
+static void mtk_vcodec_dec_child_dev_off(struct mtk_vcodec_dec_dev *vdec_dev,
+					 int hw_idx)
+{
+	struct mtk_vcodec_pm *pm;
+
+	pm = mtk_vcodec_dec_get_pm(vdec_dev, hw_idx);
+	if (pm) {
+		mtk_vcodec_dec_clock_off(pm);
+		mtk_vcodec_dec_pw_off(pm);
+	}
+
+	if (hw_idx == MTK_VDEC_LAT0) {
+		pm = mtk_vcodec_dec_get_pm(vdec_dev, MTK_VDEC_LAT_SOC);
+		if (pm) {
+			mtk_vcodec_dec_clock_off(pm);
+			mtk_vcodec_dec_pw_off(pm);
+		}
+	}
+}
+
+void mtk_vcodec_dec_enable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx)
+{
+	mutex_lock(&ctx->dev->dec_mutex[hw_idx]);
+
+	if (IS_VDEC_LAT_ARCH(ctx->dev->vdec_pdata->hw_arch) &&
+	    hw_idx == MTK_VDEC_CORE)
+		mtk_vcodec_dec_child_dev_on(ctx->dev, MTK_VDEC_LAT0);
+	mtk_vcodec_dec_child_dev_on(ctx->dev, hw_idx);
+
+	mtk_vcodec_dec_enable_irq(ctx->dev, hw_idx);
+
+	if (IS_VDEC_INNER_RACING(ctx->dev->dec_capability))
+		mtk_vcodec_load_racing_info(ctx);
+}
+EXPORT_SYMBOL_GPL(mtk_vcodec_dec_enable_hardware);
+
+void mtk_vcodec_dec_disable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx)
+{
+	if (IS_VDEC_INNER_RACING(ctx->dev->dec_capability))
+		mtk_vcodec_record_racing_info(ctx);
+
+	mtk_vcodec_dec_disable_irq(ctx->dev, hw_idx);
+
+	mtk_vcodec_dec_child_dev_off(ctx->dev, hw_idx);
+	if (IS_VDEC_LAT_ARCH(ctx->dev->vdec_pdata->hw_arch) &&
+	    hw_idx == MTK_VDEC_CORE)
+		mtk_vcodec_dec_child_dev_off(ctx->dev, MTK_VDEC_LAT0);
+
+	mutex_unlock(&ctx->dev->dec_mutex[hw_idx]);
+}
+EXPORT_SYMBOL_GPL(mtk_vcodec_dec_disable_hardware);
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h
new file mode 100644
index 0000000000..87a50d589d
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_pm.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#ifndef _MTK_VCODEC_DEC_PM_H_
+#define _MTK_VCODEC_DEC_PM_H_
+
+#include "mtk_vcodec_dec_drv.h"
+
+int mtk_vcodec_init_dec_clk(struct platform_device *pdev, struct mtk_vcodec_pm *pm);
+
+void mtk_vcodec_dec_enable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx);
+void mtk_vcodec_dec_disable_hardware(struct mtk_vcodec_dec_ctx *ctx, int hw_idx);
+
+#endif /* _MTK_VCODEC_DEC_PM_H_ */
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c
new file mode 100644
index 0000000000..11ca2c2fba
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateful.c
@@ -0,0 +1,623 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <media/v4l2-event.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "mtk_vcodec_dec.h"
+#include "mtk_vcodec_dec_pm.h"
+#include "vdec_drv_if.h"
+
+static struct mtk_video_fmt mtk_video_formats[] = {
+	{
+		.fourcc = V4L2_PIX_FMT_H264,
+		.type = MTK_FMT_DEC,
+		.num_planes = 1,
+		.flags = V4L2_FMT_FLAG_DYN_RESOLUTION,
+		.frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16,
+			     MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 },
+	},
+	{
+		.fourcc = V4L2_PIX_FMT_VP8,
+		.type = MTK_FMT_DEC,
+		.num_planes = 1,
+		.flags = V4L2_FMT_FLAG_DYN_RESOLUTION,
+		.frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16,
+			     MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 },
+	},
+	{
+		.fourcc = V4L2_PIX_FMT_VP9,
+		.type = MTK_FMT_DEC,
+		.num_planes = 1,
+		.flags = V4L2_FMT_FLAG_DYN_RESOLUTION,
+		.frmsize = { MTK_VDEC_MIN_W, MTK_VDEC_MAX_W, 16,
+			     MTK_VDEC_MIN_H, MTK_VDEC_MAX_H, 16 },
+	},
+	{
+		.fourcc = V4L2_PIX_FMT_MT21C,
+		.type = MTK_FMT_FRAME,
+		.num_planes = 2,
+	},
+};
+
+static const unsigned int num_supported_formats =
+	ARRAY_SIZE(mtk_video_formats);
+
+#define DEFAULT_OUT_FMT_IDX 0
+#define DEFAULT_CAP_FMT_IDX 3
+
+/*
+ * This function tries to clean all display buffers, the buffers will return
+ * in display order.
+ * Note the buffers returned from codec driver may still be in driver's
+ * reference list.
+ */
+static struct vb2_buffer *get_display_buffer(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_fb *disp_frame_buffer = NULL;
+	struct mtk_video_dec_buf *dstbuf;
+	struct vb2_v4l2_buffer *vb;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d]", ctx->id);
+	if (vdec_if_get_param(ctx, GET_PARAM_DISP_FRAME_BUFFER,
+			      &disp_frame_buffer)) {
+		mtk_v4l2_vdec_err(ctx, "[%d]Cannot get param : GET_PARAM_DISP_FRAME_BUFFER",
+				  ctx->id);
+		return NULL;
+	}
+
+	if (!disp_frame_buffer) {
+		mtk_v4l2_vdec_dbg(3, ctx, "No display frame buffer");
+		return NULL;
+	}
+
+	dstbuf = container_of(disp_frame_buffer, struct mtk_video_dec_buf,
+			      frame_buffer);
+	vb = &dstbuf->m2m_buf.vb;
+	mutex_lock(&ctx->lock);
+	if (dstbuf->used) {
+		mtk_v4l2_vdec_dbg(2, ctx, "[%d]status=%x queue id=%d to done_list %d",
+				  ctx->id, disp_frame_buffer->status,
+				  vb->vb2_buf.index, dstbuf->queued_in_vb2);
+
+		v4l2_m2m_buf_done(vb, VB2_BUF_STATE_DONE);
+		ctx->decoded_frame_cnt++;
+	}
+	mutex_unlock(&ctx->lock);
+	return &vb->vb2_buf;
+}
+
+/*
+ * This function tries to clean all capture buffers that are not used as
+ * reference buffers by codec driver any more
+ * In this case, we need re-queue buffer to vb2 buffer if user space
+ * already returns this buffer to v4l2 or this buffer is just the output of
+ * previous sps/pps/resolution change decode, or do nothing if user
+ * space still owns this buffer
+ */
+static struct vb2_buffer *get_free_buffer(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct mtk_video_dec_buf *dstbuf;
+	struct vdec_fb *free_frame_buffer = NULL;
+	struct vb2_v4l2_buffer *vb;
+
+	if (vdec_if_get_param(ctx, GET_PARAM_FREE_FRAME_BUFFER,
+			      &free_frame_buffer)) {
+		mtk_v4l2_vdec_err(ctx, "[%d] Error!! Cannot get param", ctx->id);
+		return NULL;
+	}
+	if (!free_frame_buffer) {
+		mtk_v4l2_vdec_dbg(3, ctx, " No free frame buffer");
+		return NULL;
+	}
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] tmp_frame_addr = 0x%p", ctx->id,
+			  free_frame_buffer);
+
+	dstbuf = container_of(free_frame_buffer, struct mtk_video_dec_buf,
+			      frame_buffer);
+	vb = &dstbuf->m2m_buf.vb;
+
+	mutex_lock(&ctx->lock);
+	if (dstbuf->used) {
+		if (dstbuf->queued_in_vb2 && dstbuf->queued_in_v4l2 &&
+		    free_frame_buffer->status == FB_ST_FREE) {
+			/*
+			 * After decode sps/pps or non-display buffer, we don't
+			 * need to return capture buffer to user space, but
+			 * just re-queue this capture buffer to vb2 queue.
+			 * This reduce overheads that dq/q unused capture
+			 * buffer. In this case, queued_in_vb2 = true.
+			 */
+			mtk_v4l2_vdec_dbg(2, ctx, "[%d]status=%x queue id=%d to rdy_queue %d",
+					  ctx->id, free_frame_buffer->status,
+					  vb->vb2_buf.index, dstbuf->queued_in_vb2);
+			v4l2_m2m_buf_queue(ctx->m2m_ctx, vb);
+		} else if (!dstbuf->queued_in_vb2 && dstbuf->queued_in_v4l2) {
+			/*
+			 * If buffer in v4l2 driver but not in vb2 queue yet,
+			 * and we get this buffer from free_list, it means
+			 * that codec driver do not use this buffer as
+			 * reference buffer anymore. We should q buffer to vb2
+			 * queue, so later work thread could get this buffer
+			 * for decode. In this case, queued_in_vb2 = false
+			 * means this buffer is not from previous decode
+			 * output.
+			 */
+			mtk_v4l2_vdec_dbg(2, ctx,
+					  "[%d]status=%x queue id=%d to rdy_queue",
+					  ctx->id, free_frame_buffer->status,
+					  vb->vb2_buf.index);
+			v4l2_m2m_buf_queue(ctx->m2m_ctx, vb);
+			dstbuf->queued_in_vb2 = true;
+		} else {
+			/*
+			 * Codec driver do not need to reference this capture
+			 * buffer and this buffer is not in v4l2 driver.
+			 * Then we don't need to do any thing, just add log when
+			 * we need to debug buffer flow.
+			 * When this buffer q from user space, it could
+			 * directly q to vb2 buffer
+			 */
+			mtk_v4l2_vdec_dbg(3, ctx, "[%d]status=%x err queue id=%d %d %d",
+					  ctx->id, free_frame_buffer->status,
+					  vb->vb2_buf.index, dstbuf->queued_in_vb2,
+					  dstbuf->queued_in_v4l2);
+		}
+		dstbuf->used = false;
+	}
+	mutex_unlock(&ctx->lock);
+	return &vb->vb2_buf;
+}
+
+static void clean_display_buffer(struct mtk_vcodec_dec_ctx *ctx)
+{
+	while (get_display_buffer(ctx))
+		;
+}
+
+static void clean_free_buffer(struct mtk_vcodec_dec_ctx *ctx)
+{
+	while (get_free_buffer(ctx))
+		;
+}
+
+static void mtk_vdec_queue_res_chg_event(struct mtk_vcodec_dec_ctx *ctx)
+{
+	static const struct v4l2_event ev_src_ch = {
+		.type = V4L2_EVENT_SOURCE_CHANGE,
+		.u.src_change.changes = V4L2_EVENT_SRC_CH_RESOLUTION,
+	};
+
+	mtk_v4l2_vdec_dbg(1, ctx, "[%d]", ctx->id);
+	v4l2_event_queue_fh(&ctx->fh, &ev_src_ch);
+}
+
+static int mtk_vdec_flush_decoder(struct mtk_vcodec_dec_ctx *ctx)
+{
+	bool res_chg;
+	int ret;
+
+	ret = vdec_if_decode(ctx, NULL, NULL, &res_chg);
+	if (ret)
+		mtk_v4l2_vdec_err(ctx, "DecodeFinal failed, ret=%d", ret);
+
+	clean_display_buffer(ctx);
+	clean_free_buffer(ctx);
+
+	return 0;
+}
+
+static void mtk_vdec_update_fmt(struct mtk_vcodec_dec_ctx *ctx,
+				unsigned int pixelformat)
+{
+	const struct mtk_video_fmt *fmt;
+	struct mtk_q_data *dst_q_data;
+	unsigned int k;
+
+	dst_q_data = &ctx->q_data[MTK_Q_DATA_DST];
+	for (k = 0; k < num_supported_formats; k++) {
+		fmt = &mtk_video_formats[k];
+		if (fmt->fourcc == pixelformat) {
+			mtk_v4l2_vdec_dbg(1, ctx, "Update cap fourcc(%d -> %d)",
+					  dst_q_data->fmt->fourcc, pixelformat);
+			dst_q_data->fmt = fmt;
+			return;
+		}
+	}
+
+	mtk_v4l2_vdec_err(ctx, "Cannot get fourcc(%d), using init value", pixelformat);
+}
+
+static int mtk_vdec_pic_info_update(struct mtk_vcodec_dec_ctx *ctx)
+{
+	unsigned int dpbsize = 0;
+	int ret;
+
+	if (vdec_if_get_param(ctx, GET_PARAM_PIC_INFO,
+			      &ctx->last_decoded_picinfo)) {
+		mtk_v4l2_vdec_err(ctx, "[%d]Error!! Cannot get param : GET_PARAM_PICTURE_INFO ERR",
+				  ctx->id);
+		return -EINVAL;
+	}
+
+	if (ctx->last_decoded_picinfo.pic_w == 0 ||
+	    ctx->last_decoded_picinfo.pic_h == 0 ||
+	    ctx->last_decoded_picinfo.buf_w == 0 ||
+	    ctx->last_decoded_picinfo.buf_h == 0) {
+		mtk_v4l2_vdec_err(ctx, "Cannot get correct pic info");
+		return -EINVAL;
+	}
+
+	if (ctx->last_decoded_picinfo.cap_fourcc != ctx->picinfo.cap_fourcc &&
+	    ctx->picinfo.cap_fourcc != 0)
+		mtk_vdec_update_fmt(ctx, ctx->picinfo.cap_fourcc);
+
+	if (ctx->last_decoded_picinfo.pic_w == ctx->picinfo.pic_w ||
+	    ctx->last_decoded_picinfo.pic_h == ctx->picinfo.pic_h)
+		return 0;
+
+	mtk_v4l2_vdec_dbg(1, ctx, "[%d]-> new(%d,%d), old(%d,%d), real(%d,%d)", ctx->id,
+			  ctx->last_decoded_picinfo.pic_w,
+			  ctx->last_decoded_picinfo.pic_h, ctx->picinfo.pic_w,
+			  ctx->picinfo.pic_h, ctx->last_decoded_picinfo.buf_w,
+			  ctx->last_decoded_picinfo.buf_h);
+
+	ret = vdec_if_get_param(ctx, GET_PARAM_DPB_SIZE, &dpbsize);
+	if (dpbsize == 0)
+		mtk_v4l2_vdec_err(ctx, "Incorrect dpb size, ret=%d", ret);
+
+	ctx->dpb_size = dpbsize;
+
+	return ret;
+}
+
+static void mtk_vdec_worker(struct work_struct *work)
+{
+	struct mtk_vcodec_dec_ctx *ctx =
+		container_of(work, struct mtk_vcodec_dec_ctx, decode_work);
+	struct mtk_vcodec_dec_dev *dev = ctx->dev;
+	struct vb2_v4l2_buffer *src_buf, *dst_buf;
+	struct mtk_vcodec_mem buf;
+	struct vdec_fb *pfb;
+	bool res_chg = false;
+	int ret;
+	struct mtk_video_dec_buf *dst_buf_info, *src_buf_info;
+
+	src_buf = v4l2_m2m_next_src_buf(ctx->m2m_ctx);
+	if (!src_buf) {
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		mtk_v4l2_vdec_dbg(1, ctx, "[%d] src_buf empty!!", ctx->id);
+		return;
+	}
+
+	dst_buf = v4l2_m2m_next_dst_buf(ctx->m2m_ctx);
+	if (!dst_buf) {
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		mtk_v4l2_vdec_dbg(1, ctx, "[%d] dst_buf empty!!", ctx->id);
+		return;
+	}
+
+	dst_buf_info =
+		container_of(dst_buf, struct mtk_video_dec_buf, m2m_buf.vb);
+
+	pfb = &dst_buf_info->frame_buffer;
+	pfb->base_y.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 0);
+	pfb->base_y.dma_addr =
+		vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 0);
+	pfb->base_y.size = ctx->picinfo.fb_sz[0];
+
+	pfb->base_c.va = vb2_plane_vaddr(&dst_buf->vb2_buf, 1);
+	pfb->base_c.dma_addr =
+		vb2_dma_contig_plane_dma_addr(&dst_buf->vb2_buf, 1);
+	pfb->base_c.size = ctx->picinfo.fb_sz[1];
+	pfb->status = 0;
+	mtk_v4l2_vdec_dbg(3, ctx, "===>[%d] vdec_if_decode() ===>", ctx->id);
+
+	mtk_v4l2_vdec_dbg(3, ctx,
+			  "id=%d Framebuf  pfb=%p VA=%p Y_DMA=%pad C_DMA=%pad Size=%zx",
+			  dst_buf->vb2_buf.index, pfb, pfb->base_y.va,
+			  &pfb->base_y.dma_addr, &pfb->base_c.dma_addr, pfb->base_y.size);
+
+	if (src_buf == &ctx->empty_flush_buf.vb) {
+		mtk_v4l2_vdec_dbg(1, ctx, "Got empty flush input buffer.");
+		src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+
+		/* update dst buf status */
+		dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx);
+		mutex_lock(&ctx->lock);
+		dst_buf_info->used = false;
+		mutex_unlock(&ctx->lock);
+
+		vdec_if_decode(ctx, NULL, NULL, &res_chg);
+		clean_display_buffer(ctx);
+		vb2_set_plane_payload(&dst_buf->vb2_buf, 0, 0);
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2)
+			vb2_set_plane_payload(&dst_buf->vb2_buf, 1, 0);
+		dst_buf->flags |= V4L2_BUF_FLAG_LAST;
+		v4l2_m2m_buf_done(dst_buf, VB2_BUF_STATE_DONE);
+		clean_free_buffer(ctx);
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		return;
+	}
+
+	src_buf_info =
+		container_of(src_buf, struct mtk_video_dec_buf, m2m_buf.vb);
+
+	buf.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0);
+	buf.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0);
+	buf.size = (size_t)src_buf->vb2_buf.planes[0].bytesused;
+	if (!buf.va) {
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		mtk_v4l2_vdec_err(ctx, "[%d] id=%d src_addr is NULL!!", ctx->id,
+				  src_buf->vb2_buf.index);
+		return;
+	}
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] Bitstream VA=%p DMA=%pad Size=%zx vb=%p",
+			  ctx->id, buf.va, &buf.dma_addr, buf.size, src_buf);
+	dst_buf->vb2_buf.timestamp = src_buf->vb2_buf.timestamp;
+	dst_buf->timecode = src_buf->timecode;
+	mutex_lock(&ctx->lock);
+	dst_buf_info->used = true;
+	mutex_unlock(&ctx->lock);
+	src_buf_info->used = true;
+
+	ret = vdec_if_decode(ctx, &buf, pfb, &res_chg);
+
+	if (ret) {
+		mtk_v4l2_vdec_err(ctx,
+				  "[%d] decode src[%d] sz=0x%zx pts=%llu dst[%d] ret=%d res_chg=%d",
+				  ctx->id, src_buf->vb2_buf.index, buf.size,
+				  src_buf->vb2_buf.timestamp, dst_buf->vb2_buf.index, ret, res_chg);
+		src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+		if (ret == -EIO) {
+			mutex_lock(&ctx->lock);
+			src_buf_info->error = true;
+			mutex_unlock(&ctx->lock);
+		}
+		v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR);
+	} else if (!res_chg) {
+		/*
+		 * we only return src buffer with VB2_BUF_STATE_DONE
+		 * when decode success without resolution change
+		 */
+		src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+		v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE);
+	}
+
+	dst_buf = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx);
+	clean_display_buffer(ctx);
+	clean_free_buffer(ctx);
+
+	if (!ret && res_chg) {
+		mtk_vdec_pic_info_update(ctx);
+		/*
+		 * On encountering a resolution change in the stream.
+		 * The driver must first process and decode all
+		 * remaining buffers from before the resolution change
+		 * point, so call flush decode here
+		 */
+		mtk_vdec_flush_decoder(ctx);
+		/*
+		 * After all buffers containing decoded frames from
+		 * before the resolution change point ready to be
+		 * dequeued on the CAPTURE queue, the driver sends a
+		 * V4L2_EVENT_SOURCE_CHANGE event for source change
+		 * type V4L2_EVENT_SRC_CH_RESOLUTION
+		 */
+		mtk_vdec_queue_res_chg_event(ctx);
+	}
+	v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+}
+
+static void vb2ops_vdec_stateful_buf_queue(struct vb2_buffer *vb)
+{
+	struct vb2_v4l2_buffer *src_buf;
+	struct mtk_vcodec_mem src_mem;
+	bool res_chg = false;
+	int ret;
+	unsigned int dpbsize = 1, i;
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	struct mtk_q_data *dst_q_data;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id,
+			  vb->vb2_queue->type, vb->index, vb);
+	/*
+	 * check if this buffer is ready to be used after decode
+	 */
+	if (vb->vb2_queue->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE) {
+		struct mtk_video_dec_buf *buf;
+
+		vb2_v4l2 = to_vb2_v4l2_buffer(vb);
+		buf = container_of(vb2_v4l2, struct mtk_video_dec_buf,
+				   m2m_buf.vb);
+		mutex_lock(&ctx->lock);
+		if (!buf->used) {
+			v4l2_m2m_buf_queue(ctx->m2m_ctx, vb2_v4l2);
+			buf->queued_in_vb2 = true;
+			buf->queued_in_v4l2 = true;
+		} else {
+			buf->queued_in_vb2 = false;
+			buf->queued_in_v4l2 = true;
+		}
+		mutex_unlock(&ctx->lock);
+		return;
+	}
+
+	v4l2_m2m_buf_queue(ctx->m2m_ctx, to_vb2_v4l2_buffer(vb));
+
+	if (ctx->state != MTK_STATE_INIT) {
+		mtk_v4l2_vdec_dbg(3, ctx, "[%d] already init driver %d", ctx->id, ctx->state);
+		return;
+	}
+
+	src_buf = v4l2_m2m_next_src_buf(ctx->m2m_ctx);
+	if (!src_buf) {
+		mtk_v4l2_vdec_err(ctx, "No src buffer");
+		return;
+	}
+
+	if (src_buf == &ctx->empty_flush_buf.vb) {
+		/* This shouldn't happen. Just in case. */
+		mtk_v4l2_vdec_err(ctx, "Invalid flush buffer.");
+		v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+		return;
+	}
+
+	src_mem.va = vb2_plane_vaddr(&src_buf->vb2_buf, 0);
+	src_mem.dma_addr = vb2_dma_contig_plane_dma_addr(&src_buf->vb2_buf, 0);
+	src_mem.size = (size_t)src_buf->vb2_buf.planes[0].bytesused;
+	mtk_v4l2_vdec_dbg(2, ctx, "[%d] buf id=%d va=%p dma=%pad size=%zx", ctx->id,
+			  src_buf->vb2_buf.index, src_mem.va, &src_mem.dma_addr, src_mem.size);
+
+	ret = vdec_if_decode(ctx, &src_mem, NULL, &res_chg);
+	if (ret || !res_chg) {
+		/*
+		 * fb == NULL means to parse SPS/PPS header or
+		 * resolution info in src_mem. Decode can fail
+		 * if there is no SPS header or picture info
+		 * in bs
+		 */
+
+		src_buf = v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+		if (ret == -EIO) {
+			mtk_v4l2_vdec_err(ctx, "[%d] Unrecoverable error in vdec_if_decode.",
+					  ctx->id);
+			ctx->state = MTK_STATE_ABORT;
+			v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_ERROR);
+		} else {
+			v4l2_m2m_buf_done(src_buf, VB2_BUF_STATE_DONE);
+		}
+		mtk_v4l2_vdec_dbg(ret ? 0 : 1, ctx,
+				  "[%d] decode() src_buf=%d, size=%zu, fail=%d, res_chg=%d",
+				  ctx->id, src_buf->vb2_buf.index, src_mem.size, ret, res_chg);
+		return;
+	}
+
+	if (vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo)) {
+		mtk_v4l2_vdec_err(ctx, "[%d]Error!! Cannot get param : GET_PARAM_PICTURE_INFO ERR",
+				  ctx->id);
+		return;
+	}
+
+	ctx->last_decoded_picinfo = ctx->picinfo;
+	dst_q_data = &ctx->q_data[MTK_Q_DATA_DST];
+	for (i = 0; i < dst_q_data->fmt->num_planes; i++) {
+		dst_q_data->sizeimage[i] = ctx->picinfo.fb_sz[i];
+		dst_q_data->bytesperline[i] = ctx->picinfo.buf_w;
+	}
+
+	mtk_v4l2_vdec_dbg(2, ctx, "[%d] init OK wxh=%dx%d pic wxh=%dx%d sz[0]=0x%x sz[1]=0x%x",
+			  ctx->id, ctx->picinfo.buf_w, ctx->picinfo.buf_h, ctx->picinfo.pic_w,
+			  ctx->picinfo.pic_h, dst_q_data->sizeimage[0], dst_q_data->sizeimage[1]);
+
+	ret = vdec_if_get_param(ctx, GET_PARAM_DPB_SIZE, &dpbsize);
+	if (dpbsize == 0)
+		mtk_v4l2_vdec_err(ctx, "[%d] GET_PARAM_DPB_SIZE fail=%d", ctx->id, ret);
+
+	ctx->dpb_size = dpbsize;
+	ctx->state = MTK_STATE_HEADER;
+	mtk_v4l2_vdec_dbg(1, ctx, "[%d] dpbsize=%d", ctx->id, ctx->dpb_size);
+
+	mtk_vdec_queue_res_chg_event(ctx);
+}
+
+static int mtk_vdec_g_v_ctrl(struct v4l2_ctrl *ctrl)
+{
+	struct mtk_vcodec_dec_ctx *ctx = ctrl_to_dec_ctx(ctrl);
+	int ret = 0;
+
+	switch (ctrl->id) {
+	case V4L2_CID_MIN_BUFFERS_FOR_CAPTURE:
+		if (ctx->state >= MTK_STATE_HEADER) {
+			ctrl->val = ctx->dpb_size;
+		} else {
+			mtk_v4l2_vdec_dbg(0, ctx, "Seqinfo not ready");
+			ctrl->val = 0;
+		}
+		break;
+	default:
+		ret = -EINVAL;
+	}
+	return ret;
+}
+
+static const struct v4l2_ctrl_ops mtk_vcodec_dec_ctrl_ops = {
+	.g_volatile_ctrl = mtk_vdec_g_v_ctrl,
+};
+
+static int mtk_vcodec_dec_ctrls_setup(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct v4l2_ctrl *ctrl;
+
+	v4l2_ctrl_handler_init(&ctx->ctrl_hdl, 1);
+
+	ctrl = v4l2_ctrl_new_std(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops,
+				 V4L2_CID_MIN_BUFFERS_FOR_CAPTURE, 0, 32, 1, 1);
+	ctrl->flags |= V4L2_CTRL_FLAG_VOLATILE;
+	v4l2_ctrl_new_std_menu(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops,
+			       V4L2_CID_MPEG_VIDEO_VP9_PROFILE,
+			       V4L2_MPEG_VIDEO_VP9_PROFILE_0, 0,
+			       V4L2_MPEG_VIDEO_VP9_PROFILE_0);
+	/*
+	 * H264. Baseline / Extended decoding is not supported.
+	 */
+	v4l2_ctrl_new_std_menu(&ctx->ctrl_hdl, &mtk_vcodec_dec_ctrl_ops,
+			       V4L2_CID_MPEG_VIDEO_H264_PROFILE, V4L2_MPEG_VIDEO_H264_PROFILE_HIGH,
+			       BIT(V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE) |
+			       BIT(V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED),
+			       V4L2_MPEG_VIDEO_H264_PROFILE_MAIN);
+
+	if (ctx->ctrl_hdl.error) {
+		mtk_v4l2_vdec_err(ctx, "Adding control failed %d", ctx->ctrl_hdl.error);
+		return ctx->ctrl_hdl.error;
+	}
+
+	v4l2_ctrl_handler_setup(&ctx->ctrl_hdl);
+	return 0;
+}
+
+static void mtk_init_vdec_params(struct mtk_vcodec_dec_ctx *ctx)
+{
+	unsigned int i;
+
+	if (!(ctx->dev->dec_capability & VCODEC_CAPABILITY_4K_DISABLED)) {
+		for (i = 0; i < num_supported_formats; i++) {
+			if (mtk_video_formats[i].type != MTK_FMT_DEC)
+				continue;
+
+			mtk_video_formats[i].frmsize.max_width =
+				VCODEC_DEC_4K_CODED_WIDTH;
+			mtk_video_formats[i].frmsize.max_height =
+				VCODEC_DEC_4K_CODED_HEIGHT;
+		}
+	}
+}
+
+static struct vb2_ops mtk_vdec_frame_vb2_ops = {
+	.queue_setup = vb2ops_vdec_queue_setup,
+	.buf_prepare = vb2ops_vdec_buf_prepare,
+	.wait_prepare = vb2_ops_wait_prepare,
+	.wait_finish = vb2_ops_wait_finish,
+	.start_streaming = vb2ops_vdec_start_streaming,
+
+	.buf_queue = vb2ops_vdec_stateful_buf_queue,
+	.buf_init = vb2ops_vdec_buf_init,
+	.buf_finish = vb2ops_vdec_buf_finish,
+	.stop_streaming = vb2ops_vdec_stop_streaming,
+};
+
+const struct mtk_vcodec_dec_pdata mtk_vdec_8173_pdata = {
+	.init_vdec_params = mtk_init_vdec_params,
+	.ctrls_setup = mtk_vcodec_dec_ctrls_setup,
+	.vdec_vb2_ops = &mtk_vdec_frame_vb2_ops,
+	.vdec_formats = mtk_video_formats,
+	.num_formats = &num_supported_formats,
+	.default_out_fmt = &mtk_video_formats[DEFAULT_OUT_FMT_IDX],
+	.default_cap_fmt = &mtk_video_formats[DEFAULT_CAP_FMT_IDX],
+	.worker = mtk_vdec_worker,
+	.flush_decoder = mtk_vdec_flush_decoder,
+	.is_subdev_supported = false,
+	.hw_arch = MTK_VDEC_PURE_SINGLE_CORE,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c
new file mode 100644
index 0000000000..e29c9c58f3
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/mtk_vcodec_dec_stateless.c
@@ -0,0 +1,756 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <media/videobuf2-v4l2.h>
+#include <media/videobuf2-dma-contig.h>
+#include <media/v4l2-event.h>
+#include <media/v4l2-mem2mem.h>
+#include <linux/module.h>
+
+#include "mtk_vcodec_dec.h"
+#include "mtk_vcodec_dec_pm.h"
+#include "vdec_drv_if.h"
+
+/**
+ * struct mtk_stateless_control  - CID control type
+ * @cfg: control configuration
+ * @codec_type: codec type (V4L2 pixel format) for CID control type
+ */
+struct mtk_stateless_control {
+	struct v4l2_ctrl_config cfg;
+	int codec_type;
+};
+
+static const struct mtk_stateless_control mtk_stateless_controls[] = {
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_SPS,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_PPS,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_SCALING_MATRIX,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_DECODE_PARAMS,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_H264_PROFILE,
+			.def = V4L2_MPEG_VIDEO_H264_PROFILE_MAIN,
+			.max = V4L2_MPEG_VIDEO_H264_PROFILE_HIGH,
+			.menu_skip_mask =
+				BIT(V4L2_MPEG_VIDEO_H264_PROFILE_BASELINE) |
+				BIT(V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED),
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_DECODE_MODE,
+			.min = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+			.def = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+			.max = V4L2_STATELESS_H264_DECODE_MODE_FRAME_BASED,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_H264_START_CODE,
+			.min = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+			.def = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+			.max = V4L2_STATELESS_H264_START_CODE_ANNEX_B,
+		},
+		.codec_type = V4L2_PIX_FMT_H264_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_VP8_FRAME,
+		},
+		.codec_type = V4L2_PIX_FMT_VP8_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_VP8_PROFILE,
+			.min = V4L2_MPEG_VIDEO_VP8_PROFILE_0,
+			.def = V4L2_MPEG_VIDEO_VP8_PROFILE_0,
+			.max = V4L2_MPEG_VIDEO_VP8_PROFILE_3,
+		},
+		.codec_type = V4L2_PIX_FMT_VP8_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_VP9_FRAME,
+		},
+		.codec_type = V4L2_PIX_FMT_VP9_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_VP9_PROFILE,
+			.min = V4L2_MPEG_VIDEO_VP9_PROFILE_0,
+			.def = V4L2_MPEG_VIDEO_VP9_PROFILE_0,
+			.max = V4L2_MPEG_VIDEO_VP9_PROFILE_3,
+		},
+		.codec_type = V4L2_PIX_FMT_VP9_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_SPS,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_PPS,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_SCALING_MATRIX,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_DECODE_PARAMS,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_HEVC_PROFILE,
+			.def = V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN,
+			.max = V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN_10,
+			.menu_skip_mask =
+				BIT(V4L2_MPEG_VIDEO_HEVC_PROFILE_MAIN_STILL_PICTURE),
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_DECODE_MODE,
+			.min = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED,
+			.def = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED,
+			.max = V4L2_STATELESS_HEVC_DECODE_MODE_FRAME_BASED,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_HEVC_START_CODE,
+			.min = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B,
+			.def = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B,
+			.max = V4L2_STATELESS_HEVC_START_CODE_ANNEX_B,
+		},
+		.codec_type = V4L2_PIX_FMT_HEVC_SLICE,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_AV1_SEQUENCE,
+
+		},
+		.codec_type = V4L2_PIX_FMT_AV1_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_AV1_FRAME,
+
+		},
+		.codec_type = V4L2_PIX_FMT_AV1_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_STATELESS_AV1_TILE_GROUP_ENTRY,
+			.dims = { V4L2_AV1_MAX_TILE_COUNT },
+
+		},
+		.codec_type = V4L2_PIX_FMT_AV1_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_AV1_PROFILE,
+			.min = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN,
+			.def = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN,
+			.max = V4L2_MPEG_VIDEO_AV1_PROFILE_MAIN,
+		},
+		.codec_type = V4L2_PIX_FMT_AV1_FRAME,
+	},
+	{
+		.cfg = {
+			.id = V4L2_CID_MPEG_VIDEO_AV1_LEVEL,
+			.min = V4L2_MPEG_VIDEO_AV1_LEVEL_2_0,
+			.def = V4L2_MPEG_VIDEO_AV1_LEVEL_4_0,
+			.max = V4L2_MPEG_VIDEO_AV1_LEVEL_5_1,
+		},
+		.codec_type = V4L2_PIX_FMT_AV1_FRAME,
+	},
+};
+
+#define NUM_CTRLS ARRAY_SIZE(mtk_stateless_controls)
+
+static struct mtk_video_fmt mtk_video_formats[9];
+
+static struct mtk_video_fmt default_out_format;
+static struct mtk_video_fmt default_cap_format;
+static unsigned int num_formats;
+
+static const struct v4l2_frmsize_stepwise stepwise_fhd = {
+	.min_width = MTK_VDEC_MIN_W,
+	.max_width = MTK_VDEC_MAX_W,
+	.step_width = 16,
+	.min_height = MTK_VDEC_MIN_H,
+	.max_height = MTK_VDEC_MAX_H,
+	.step_height = 16
+};
+
+static void mtk_vdec_stateless_cap_to_disp(struct mtk_vcodec_dec_ctx *ctx, int error,
+					   struct media_request *src_buf_req)
+{
+	struct vb2_v4l2_buffer *vb2_dst;
+	enum vb2_buffer_state state;
+
+	if (error)
+		state = VB2_BUF_STATE_ERROR;
+	else
+		state = VB2_BUF_STATE_DONE;
+
+	vb2_dst = v4l2_m2m_dst_buf_remove(ctx->m2m_ctx);
+	if (vb2_dst) {
+		v4l2_m2m_buf_done(vb2_dst, state);
+		mtk_v4l2_vdec_dbg(2, ctx, "free frame buffer id:%d to done list",
+				  vb2_dst->vb2_buf.index);
+	} else {
+		mtk_v4l2_vdec_err(ctx, "dst buffer is NULL");
+	}
+
+	if (src_buf_req)
+		v4l2_ctrl_request_complete(src_buf_req, &ctx->ctrl_hdl);
+}
+
+static struct vdec_fb *vdec_get_cap_buffer(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct mtk_video_dec_buf *framebuf;
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	struct vb2_buffer *dst_buf;
+	struct vdec_fb *pfb;
+
+	vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx);
+	if (!vb2_v4l2) {
+		mtk_v4l2_vdec_dbg(1, ctx, "[%d] dst_buf empty!!", ctx->id);
+		return NULL;
+	}
+
+	dst_buf = &vb2_v4l2->vb2_buf;
+	framebuf = container_of(vb2_v4l2, struct mtk_video_dec_buf, m2m_buf.vb);
+
+	pfb = &framebuf->frame_buffer;
+	pfb->base_y.va = vb2_plane_vaddr(dst_buf, 0);
+	pfb->base_y.dma_addr = vb2_dma_contig_plane_dma_addr(dst_buf, 0);
+	pfb->base_y.size = ctx->q_data[MTK_Q_DATA_DST].sizeimage[0];
+
+	if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2) {
+		pfb->base_c.va = vb2_plane_vaddr(dst_buf, 1);
+		pfb->base_c.dma_addr =
+			vb2_dma_contig_plane_dma_addr(dst_buf, 1);
+		pfb->base_c.size = ctx->q_data[MTK_Q_DATA_DST].sizeimage[1];
+	}
+	mtk_v4l2_vdec_dbg(1, ctx,
+			  "id=%d Framebuf pfb=%p VA=%p Y/C_DMA=%pad_%pad Sz=%zx frame_count = %d",
+			  dst_buf->index, pfb, pfb->base_y.va, &pfb->base_y.dma_addr,
+			  &pfb->base_c.dma_addr, pfb->base_y.size, ctx->decoded_frame_cnt);
+
+	return pfb;
+}
+
+static void vb2ops_vdec_buf_request_complete(struct vb2_buffer *vb)
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+
+	v4l2_ctrl_request_complete(vb->req_obj.req, &ctx->ctrl_hdl);
+}
+
+static void mtk_vdec_worker(struct work_struct *work)
+{
+	struct mtk_vcodec_dec_ctx *ctx =
+		container_of(work, struct mtk_vcodec_dec_ctx, decode_work);
+	struct mtk_vcodec_dec_dev *dev = ctx->dev;
+	struct vb2_v4l2_buffer *vb2_v4l2_src;
+	struct vb2_buffer *vb2_src;
+	struct mtk_vcodec_mem *bs_src;
+	struct mtk_video_dec_buf *dec_buf_src;
+	struct media_request *src_buf_req;
+	enum vb2_buffer_state state;
+	bool res_chg = false;
+	int ret;
+
+	vb2_v4l2_src = v4l2_m2m_next_src_buf(ctx->m2m_ctx);
+	if (!vb2_v4l2_src) {
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		mtk_v4l2_vdec_dbg(1, ctx, "[%d] no available source buffer", ctx->id);
+		return;
+	}
+
+	vb2_src = &vb2_v4l2_src->vb2_buf;
+	dec_buf_src = container_of(vb2_v4l2_src, struct mtk_video_dec_buf,
+				   m2m_buf.vb);
+	bs_src = &dec_buf_src->bs_buffer;
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id,
+			  vb2_src->vb2_queue->type, vb2_src->index, vb2_src);
+
+	bs_src->va = vb2_plane_vaddr(vb2_src, 0);
+	bs_src->dma_addr = vb2_dma_contig_plane_dma_addr(vb2_src, 0);
+	bs_src->size = (size_t)vb2_src->planes[0].bytesused;
+	if (!bs_src->va) {
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+		mtk_v4l2_vdec_err(ctx, "[%d] id=%d source buffer is NULL", ctx->id,
+				  vb2_src->index);
+		return;
+	}
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] Bitstream VA=%p DMA=%pad Size=%zx vb=%p",
+			  ctx->id, bs_src->va, &bs_src->dma_addr, bs_src->size, vb2_src);
+	/* Apply request controls. */
+	src_buf_req = vb2_src->req_obj.req;
+	if (src_buf_req)
+		v4l2_ctrl_request_setup(src_buf_req, &ctx->ctrl_hdl);
+	else
+		mtk_v4l2_vdec_err(ctx, "vb2 buffer media request is NULL");
+
+	ret = vdec_if_decode(ctx, bs_src, NULL, &res_chg);
+	if (ret && ret != -EAGAIN) {
+		mtk_v4l2_vdec_err(ctx,
+				  "[%d] decode src_buf[%d] sz=0x%zx pts=%llu ret=%d res_chg=%d",
+				  ctx->id, vb2_src->index, bs_src->size,
+				  vb2_src->timestamp, ret, res_chg);
+		if (ret == -EIO) {
+			mutex_lock(&ctx->lock);
+			dec_buf_src->error = true;
+			mutex_unlock(&ctx->lock);
+		}
+	}
+
+	state = ret ? VB2_BUF_STATE_ERROR : VB2_BUF_STATE_DONE;
+	if (!IS_VDEC_LAT_ARCH(dev->vdec_pdata->hw_arch) ||
+	    ctx->current_codec == V4L2_PIX_FMT_VP8_FRAME) {
+		v4l2_m2m_buf_done_and_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx, state);
+		if (src_buf_req)
+			v4l2_ctrl_request_complete(src_buf_req, &ctx->ctrl_hdl);
+	} else {
+		if (ret != -EAGAIN) {
+			v4l2_m2m_src_buf_remove(ctx->m2m_ctx);
+			v4l2_m2m_buf_done(vb2_v4l2_src, state);
+		}
+		v4l2_m2m_job_finish(dev->m2m_dev_dec, ctx->m2m_ctx);
+	}
+}
+
+static void vb2ops_vdec_stateless_buf_queue(struct vb2_buffer *vb)
+{
+	struct mtk_vcodec_dec_ctx *ctx = vb2_get_drv_priv(vb->vb2_queue);
+	struct vb2_v4l2_buffer *vb2_v4l2 = to_vb2_v4l2_buffer(vb);
+
+	mtk_v4l2_vdec_dbg(3, ctx, "[%d] (%d) id=%d, vb=%p", ctx->id, vb->vb2_queue->type,
+			  vb->index, vb);
+
+	mutex_lock(&ctx->lock);
+	v4l2_m2m_buf_queue(ctx->m2m_ctx, vb2_v4l2);
+	mutex_unlock(&ctx->lock);
+	if (vb->vb2_queue->type != V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE)
+		return;
+
+	/* If an OUTPUT buffer, we may need to update the state */
+	if (ctx->state == MTK_STATE_INIT) {
+		ctx->state = MTK_STATE_HEADER;
+		mtk_v4l2_vdec_dbg(1, ctx, "Init driver from init to header.");
+	} else {
+		mtk_v4l2_vdec_dbg(3, ctx, "[%d] already init driver %d", ctx->id, ctx->state);
+	}
+}
+
+static int mtk_vdec_flush_decoder(struct mtk_vcodec_dec_ctx *ctx)
+{
+	bool res_chg;
+
+	return vdec_if_decode(ctx, NULL, NULL, &res_chg);
+}
+
+static int mtk_vcodec_get_pic_info(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct mtk_q_data *q_data;
+	int ret = 0;
+
+	q_data = &ctx->q_data[MTK_Q_DATA_DST];
+	if (q_data->fmt->num_planes == 1) {
+		mtk_v4l2_vdec_err(ctx, "[%d]Error!! 10bit mode not support one plane", ctx->id);
+		return -EINVAL;
+	}
+
+	ctx->capture_fourcc = q_data->fmt->fourcc;
+	ret = vdec_if_get_param(ctx, GET_PARAM_PIC_INFO, &ctx->picinfo);
+	if (ret) {
+		mtk_v4l2_vdec_err(ctx, "[%d]Error!! Get GET_PARAM_PICTURE_INFO Fail", ctx->id);
+		return ret;
+	}
+
+	ctx->last_decoded_picinfo = ctx->picinfo;
+
+	q_data->sizeimage[0] = ctx->picinfo.fb_sz[0];
+	q_data->bytesperline[0] = ctx->picinfo.buf_w * 5 / 4;
+
+	q_data->sizeimage[1] = ctx->picinfo.fb_sz[1];
+	q_data->bytesperline[1] = ctx->picinfo.buf_w * 5 / 4;
+
+	q_data->coded_width = ctx->picinfo.buf_w;
+	q_data->coded_height = ctx->picinfo.buf_h;
+	mtk_v4l2_vdec_dbg(1, ctx, "[%d] wxh=%dx%d pic wxh=%dx%d sz[0]=0x%x sz[1]=0x%x",
+			  ctx->id, ctx->picinfo.buf_w, ctx->picinfo.buf_h,
+			  ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+			  q_data->sizeimage[0], q_data->sizeimage[1]);
+
+	return ret;
+}
+
+static int mtk_vdec_s_ctrl(struct v4l2_ctrl *ctrl)
+{
+	struct mtk_vcodec_dec_ctx *ctx = ctrl_to_dec_ctx(ctrl);
+	struct v4l2_ctrl_h264_sps *h264;
+	struct v4l2_ctrl_hevc_sps *h265;
+	struct v4l2_ctrl_vp9_frame *frame;
+	struct v4l2_ctrl_av1_sequence *seq;
+	struct v4l2_ctrl *hdr_ctrl;
+	const struct mtk_vcodec_dec_pdata *dec_pdata = ctx->dev->vdec_pdata;
+	const struct mtk_video_fmt *fmt;
+	int i = 0, ret = 0;
+
+	hdr_ctrl = ctrl;
+	if (!hdr_ctrl || !hdr_ctrl->p_new.p)
+		return -EINVAL;
+
+	switch (hdr_ctrl->id) {
+	case V4L2_CID_STATELESS_H264_SPS:
+		h264 = (struct v4l2_ctrl_h264_sps *)hdr_ctrl->p_new.p;
+
+		if (h264->bit_depth_chroma_minus8 == 2 && h264->bit_depth_luma_minus8 == 2) {
+			ctx->is_10bit_bitstream = true;
+		} else if (h264->bit_depth_chroma_minus8 != 0 &&
+			   h264->bit_depth_luma_minus8 != 0) {
+			mtk_v4l2_vdec_err(ctx, "H264: chroma_minus8:%d, luma_minus8:%d",
+					  h264->bit_depth_chroma_minus8,
+					  h264->bit_depth_luma_minus8);
+			return -EINVAL;
+		}
+		break;
+	case V4L2_CID_STATELESS_HEVC_SPS:
+		h265 = (struct v4l2_ctrl_hevc_sps *)hdr_ctrl->p_new.p;
+
+		if (h265->bit_depth_chroma_minus8 == 2 && h265->bit_depth_luma_minus8 == 2) {
+			ctx->is_10bit_bitstream = true;
+		} else if (h265->bit_depth_chroma_minus8 != 0 &&
+			   h265->bit_depth_luma_minus8 != 0) {
+			mtk_v4l2_vdec_err(ctx, "HEVC: chroma_minus8:%d, luma_minus8:%d",
+					  h265->bit_depth_chroma_minus8,
+					  h265->bit_depth_luma_minus8);
+			return -EINVAL;
+		}
+		break;
+	case V4L2_CID_STATELESS_VP9_FRAME:
+		frame = (struct v4l2_ctrl_vp9_frame *)hdr_ctrl->p_new.p;
+
+		if (frame->bit_depth == 10) {
+			ctx->is_10bit_bitstream = true;
+		} else if (frame->bit_depth != 8) {
+			mtk_v4l2_vdec_err(ctx, "VP9: bit_depth:%d", frame->bit_depth);
+			return -EINVAL;
+		}
+		break;
+	case V4L2_CID_STATELESS_AV1_SEQUENCE:
+		seq = (struct v4l2_ctrl_av1_sequence *)hdr_ctrl->p_new.p;
+
+		if (seq->bit_depth == 10) {
+			ctx->is_10bit_bitstream = true;
+		} else if (seq->bit_depth != 8) {
+			mtk_v4l2_vdec_err(ctx, "AV1: bit_depth:%d", seq->bit_depth);
+			return -EINVAL;
+		}
+		break;
+	default:
+		mtk_v4l2_vdec_dbg(3, ctx, "Not supported to set ctrl id: 0x%x\n", hdr_ctrl->id);
+		return ret;
+	}
+
+	if (!ctx->is_10bit_bitstream)
+		return ret;
+
+	for (i = 0; i < *dec_pdata->num_formats; i++) {
+		fmt = &dec_pdata->vdec_formats[i];
+		if (fmt->fourcc == V4L2_PIX_FMT_MT2110R &&
+		    hdr_ctrl->id == V4L2_CID_STATELESS_H264_SPS) {
+			ctx->q_data[MTK_Q_DATA_DST].fmt = fmt;
+			break;
+		}
+
+		if (fmt->fourcc == V4L2_PIX_FMT_MT2110T &&
+		    (hdr_ctrl->id == V4L2_CID_STATELESS_HEVC_SPS ||
+		    hdr_ctrl->id == V4L2_CID_STATELESS_VP9_FRAME ||
+		    hdr_ctrl->id == V4L2_CID_STATELESS_AV1_SEQUENCE)) {
+			ctx->q_data[MTK_Q_DATA_DST].fmt = fmt;
+			break;
+		}
+	}
+	ret = mtk_vcodec_get_pic_info(ctx);
+
+	return ret;
+}
+
+static const struct v4l2_ctrl_ops mtk_vcodec_dec_ctrl_ops = {
+	.s_ctrl = mtk_vdec_s_ctrl,
+};
+
+static int mtk_vcodec_dec_ctrls_setup(struct mtk_vcodec_dec_ctx *ctx)
+{
+	unsigned int i;
+
+	v4l2_ctrl_handler_init(&ctx->ctrl_hdl, NUM_CTRLS);
+	if (ctx->ctrl_hdl.error) {
+		mtk_v4l2_vdec_err(ctx, "v4l2_ctrl_handler_init failed\n");
+		return ctx->ctrl_hdl.error;
+	}
+
+	for (i = 0; i < NUM_CTRLS; i++) {
+		struct v4l2_ctrl_config cfg = mtk_stateless_controls[i].cfg;
+		cfg.ops = &mtk_vcodec_dec_ctrl_ops;
+		v4l2_ctrl_new_custom(&ctx->ctrl_hdl, &cfg, NULL);
+		if (ctx->ctrl_hdl.error) {
+			mtk_v4l2_vdec_err(ctx, "Adding control %d failed %d", i,
+					  ctx->ctrl_hdl.error);
+			return ctx->ctrl_hdl.error;
+		}
+	}
+
+	v4l2_ctrl_handler_setup(&ctx->ctrl_hdl);
+
+	return 0;
+}
+
+static int fops_media_request_validate(struct media_request *mreq)
+{
+	const unsigned int buffer_cnt = vb2_request_buffer_cnt(mreq);
+
+	switch (buffer_cnt) {
+	case 1:
+		/* We expect exactly one buffer with the request */
+		break;
+	case 0:
+		pr_debug(MTK_DBG_VCODEC_STR "No buffer provided with the request.");
+		return -ENOENT;
+	default:
+		pr_debug(MTK_DBG_VCODEC_STR "Too many buffers (%d) provided with the request.",
+			 buffer_cnt);
+		return -EINVAL;
+	}
+
+	return vb2_request_validate(mreq);
+}
+
+const struct media_device_ops mtk_vcodec_media_ops = {
+	.req_validate	= fops_media_request_validate,
+	.req_queue	= v4l2_m2m_request_queue,
+};
+
+static void mtk_vcodec_add_formats(unsigned int fourcc,
+				   struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct mtk_vcodec_dec_dev *dev = ctx->dev;
+	const struct mtk_vcodec_dec_pdata *pdata = dev->vdec_pdata;
+	int count_formats = *pdata->num_formats;
+
+	switch (fourcc) {
+	case V4L2_PIX_FMT_H264_SLICE:
+	case V4L2_PIX_FMT_VP8_FRAME:
+	case V4L2_PIX_FMT_VP9_FRAME:
+	case V4L2_PIX_FMT_HEVC_SLICE:
+	case V4L2_PIX_FMT_AV1_FRAME:
+		mtk_video_formats[count_formats].fourcc = fourcc;
+		mtk_video_formats[count_formats].type = MTK_FMT_DEC;
+		mtk_video_formats[count_formats].num_planes = 1;
+		mtk_video_formats[count_formats].frmsize = stepwise_fhd;
+
+		if (!(ctx->dev->dec_capability & VCODEC_CAPABILITY_4K_DISABLED) &&
+		    fourcc != V4L2_PIX_FMT_VP8_FRAME) {
+			mtk_video_formats[count_formats].frmsize.max_width =
+				VCODEC_DEC_4K_CODED_WIDTH;
+			mtk_video_formats[count_formats].frmsize.max_height =
+				VCODEC_DEC_4K_CODED_HEIGHT;
+		}
+		break;
+	case V4L2_PIX_FMT_MM21:
+	case V4L2_PIX_FMT_MT21C:
+	case V4L2_PIX_FMT_MT2110T:
+	case V4L2_PIX_FMT_MT2110R:
+		mtk_video_formats[count_formats].fourcc = fourcc;
+		mtk_video_formats[count_formats].type = MTK_FMT_FRAME;
+		mtk_video_formats[count_formats].num_planes = 2;
+		break;
+	default:
+		mtk_v4l2_vdec_err(ctx, "Can not add unsupported format type");
+		return;
+	}
+
+	num_formats++;
+	mtk_v4l2_vdec_dbg(3, ctx, "num_formats: %d dec_capability: 0x%x",
+			  count_formats, ctx->dev->dec_capability);
+}
+
+static void mtk_vcodec_get_supported_formats(struct mtk_vcodec_dec_ctx *ctx)
+{
+	int cap_format_count = 0, out_format_count = 0;
+
+	if (num_formats)
+		return;
+
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_MT21C) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_MT21C, ctx);
+		cap_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_IS_SUPPORT_10BIT) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_MT2110T, ctx);
+		cap_format_count++;
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_MT2110R, ctx);
+		cap_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_MM21) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_MM21, ctx);
+		cap_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_H264_SLICE) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_H264_SLICE, ctx);
+		out_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_VP8_FRAME) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_VP8_FRAME, ctx);
+		out_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_VP9_FRAME) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_VP9_FRAME, ctx);
+		out_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_HEVC_FRAME) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_HEVC_SLICE, ctx);
+		out_format_count++;
+	}
+	if (ctx->dev->dec_capability & MTK_VDEC_FORMAT_AV1_FRAME) {
+		mtk_vcodec_add_formats(V4L2_PIX_FMT_AV1_FRAME, ctx);
+		out_format_count++;
+	}
+
+	if (cap_format_count)
+		default_cap_format = mtk_video_formats[cap_format_count - 1];
+	if (out_format_count)
+		default_out_format =
+			mtk_video_formats[cap_format_count + out_format_count - 1];
+}
+
+static void mtk_init_vdec_params(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vb2_queue *src_vq;
+
+	src_vq = v4l2_m2m_get_vq(ctx->m2m_ctx,
+				 V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE);
+
+	if (!ctx->dev->vdec_pdata->is_subdev_supported)
+		ctx->dev->dec_capability |=
+			MTK_VDEC_FORMAT_H264_SLICE | MTK_VDEC_FORMAT_MM21;
+	mtk_vcodec_get_supported_formats(ctx);
+
+	/* Support request api for output plane */
+	src_vq->supports_requests = true;
+	src_vq->requires_requests = true;
+}
+
+static int vb2ops_vdec_out_buf_validate(struct vb2_buffer *vb)
+{
+	struct vb2_v4l2_buffer *vbuf = to_vb2_v4l2_buffer(vb);
+
+	vbuf->field = V4L2_FIELD_NONE;
+	return 0;
+}
+
+static struct vb2_ops mtk_vdec_request_vb2_ops = {
+	.queue_setup	= vb2ops_vdec_queue_setup,
+	.wait_prepare	= vb2_ops_wait_prepare,
+	.wait_finish	= vb2_ops_wait_finish,
+	.start_streaming	= vb2ops_vdec_start_streaming,
+	.stop_streaming	= vb2ops_vdec_stop_streaming,
+
+	.buf_queue	= vb2ops_vdec_stateless_buf_queue,
+	.buf_out_validate = vb2ops_vdec_out_buf_validate,
+	.buf_init	= vb2ops_vdec_buf_init,
+	.buf_prepare	= vb2ops_vdec_buf_prepare,
+	.buf_finish	= vb2ops_vdec_buf_finish,
+	.buf_request_complete = vb2ops_vdec_buf_request_complete,
+};
+
+const struct mtk_vcodec_dec_pdata mtk_vdec_8183_pdata = {
+	.init_vdec_params = mtk_init_vdec_params,
+	.ctrls_setup = mtk_vcodec_dec_ctrls_setup,
+	.vdec_vb2_ops = &mtk_vdec_request_vb2_ops,
+	.vdec_formats = mtk_video_formats,
+	.num_formats = &num_formats,
+	.default_out_fmt = &default_out_format,
+	.default_cap_fmt = &default_cap_format,
+	.uses_stateless_api = true,
+	.worker = mtk_vdec_worker,
+	.flush_decoder = mtk_vdec_flush_decoder,
+	.cap_to_disp = mtk_vdec_stateless_cap_to_disp,
+	.get_cap_buffer = vdec_get_cap_buffer,
+	.is_subdev_supported = false,
+	.hw_arch = MTK_VDEC_PURE_SINGLE_CORE,
+};
+
+/* This platform data is used for one lat and one core architecture. */
+const struct mtk_vcodec_dec_pdata mtk_lat_sig_core_pdata = {
+	.init_vdec_params = mtk_init_vdec_params,
+	.ctrls_setup = mtk_vcodec_dec_ctrls_setup,
+	.vdec_vb2_ops = &mtk_vdec_request_vb2_ops,
+	.vdec_formats = mtk_video_formats,
+	.num_formats = &num_formats,
+	.default_out_fmt = &default_out_format,
+	.default_cap_fmt = &default_cap_format,
+	.uses_stateless_api = true,
+	.worker = mtk_vdec_worker,
+	.flush_decoder = mtk_vdec_flush_decoder,
+	.cap_to_disp = mtk_vdec_stateless_cap_to_disp,
+	.get_cap_buffer = vdec_get_cap_buffer,
+	.is_subdev_supported = true,
+	.hw_arch = MTK_VDEC_LAT_SINGLE_CORE,
+};
+
+const struct mtk_vcodec_dec_pdata mtk_vdec_single_core_pdata = {
+	.init_vdec_params = mtk_init_vdec_params,
+	.ctrls_setup = mtk_vcodec_dec_ctrls_setup,
+	.vdec_vb2_ops = &mtk_vdec_request_vb2_ops,
+	.vdec_formats = mtk_video_formats,
+	.num_formats = &num_formats,
+	.default_out_fmt = &default_out_format,
+	.default_cap_fmt = &default_cap_format,
+	.uses_stateless_api = true,
+	.worker = mtk_vdec_worker,
+	.flush_decoder = mtk_vdec_flush_decoder,
+	.cap_to_disp = mtk_vdec_stateless_cap_to_disp,
+	.get_cap_buffer = vdec_get_cap_buffer,
+	.is_subdev_supported = true,
+	.hw_arch = MTK_VDEC_PURE_SINGLE_CORE,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c
new file mode 100644
index 0000000000..2b6a5adbc4
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_av1_req_lat_if.c
@@ -0,0 +1,2209 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2023 MediaTek Inc.
+ * Author: Xiaoyong Lu <xiaoyong.lu@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+
+#define AV1_MAX_FRAME_BUF_COUNT		(V4L2_AV1_TOTAL_REFS_PER_FRAME + 1)
+#define AV1_TILE_BUF_SIZE		64
+#define AV1_SCALE_SUBPEL_BITS		10
+#define AV1_REF_SCALE_SHIFT		14
+#define AV1_REF_NO_SCALE		BIT(AV1_REF_SCALE_SHIFT)
+#define AV1_REF_INVALID_SCALE		-1
+#define AV1_CDF_TABLE_BUFFER_SIZE	16384
+#define AV1_PRIMARY_REF_NONE		7
+
+#define AV1_INVALID_IDX			-1
+
+#define AV1_DIV_ROUND_UP_POW2(value, n)			\
+({							\
+	typeof(n) _n  = n;				\
+	typeof(value) _value = value;			\
+	(_value + (BIT(_n) >> 1)) >> _n;		\
+})
+
+#define AV1_DIV_ROUND_UP_POW2_SIGNED(value, n)				\
+({									\
+	typeof(n) _n_  = n;						\
+	typeof(value) _value_ = value;					\
+	(((_value_) < 0) ? -AV1_DIV_ROUND_UP_POW2(-(_value_), (_n_))	\
+		: AV1_DIV_ROUND_UP_POW2((_value_), (_n_)));		\
+})
+
+#define BIT_FLAG(x, bit)		(!!((x)->flags & (bit)))
+#define SEGMENTATION_FLAG(x, name)	(!!((x)->flags & V4L2_AV1_SEGMENTATION_FLAG_##name))
+#define QUANT_FLAG(x, name)		(!!((x)->flags & V4L2_AV1_QUANTIZATION_FLAG_##name))
+#define SEQUENCE_FLAG(x, name)		(!!((x)->flags & V4L2_AV1_SEQUENCE_FLAG_##name))
+#define FH_FLAG(x, name)		(!!((x)->flags & V4L2_AV1_FRAME_FLAG_##name))
+
+#define MINQ 0
+#define MAXQ 255
+
+#define DIV_LUT_PREC_BITS 14
+#define DIV_LUT_BITS 8
+#define DIV_LUT_NUM BIT(DIV_LUT_BITS)
+#define WARP_PARAM_REDUCE_BITS 6
+#define WARPEDMODEL_PREC_BITS 16
+
+#define SEG_LVL_ALT_Q 0
+#define SECONDARY_FILTER_STRENGTH_NUM_BITS 2
+
+static const short div_lut[DIV_LUT_NUM + 1] = {
+	16384, 16320, 16257, 16194, 16132, 16070, 16009, 15948, 15888, 15828, 15768,
+	15709, 15650, 15592, 15534, 15477, 15420, 15364, 15308, 15252, 15197, 15142,
+	15087, 15033, 14980, 14926, 14873, 14821, 14769, 14717, 14665, 14614, 14564,
+	14513, 14463, 14413, 14364, 14315, 14266, 14218, 14170, 14122, 14075, 14028,
+	13981, 13935, 13888, 13843, 13797, 13752, 13707, 13662, 13618, 13574, 13530,
+	13487, 13443, 13400, 13358, 13315, 13273, 13231, 13190, 13148, 13107, 13066,
+	13026, 12985, 12945, 12906, 12866, 12827, 12788, 12749, 12710, 12672, 12633,
+	12596, 12558, 12520, 12483, 12446, 12409, 12373, 12336, 12300, 12264, 12228,
+	12193, 12157, 12122, 12087, 12053, 12018, 11984, 11950, 11916, 11882, 11848,
+	11815, 11782, 11749, 11716, 11683, 11651, 11619, 11586, 11555, 11523, 11491,
+	11460, 11429, 11398, 11367, 11336, 11305, 11275, 11245, 11215, 11185, 11155,
+	11125, 11096, 11067, 11038, 11009, 10980, 10951, 10923, 10894, 10866, 10838,
+	10810, 10782, 10755, 10727, 10700, 10673, 10645, 10618, 10592, 10565, 10538,
+	10512, 10486, 10460, 10434, 10408, 10382, 10356, 10331, 10305, 10280, 10255,
+	10230, 10205, 10180, 10156, 10131, 10107, 10082, 10058, 10034, 10010, 9986,
+	9963,  9939,  9916,  9892,  9869,  9846,  9823,  9800,  9777,  9754,  9732,
+	9709,  9687,  9664,  9642,  9620,  9598,  9576,  9554,  9533,  9511,  9489,
+	9468,  9447,  9425,  9404,  9383,  9362,  9341,  9321,  9300,  9279,  9259,
+	9239,  9218,  9198,  9178,  9158,  9138,  9118,  9098,  9079,  9059,  9039,
+	9020,  9001,  8981,  8962,  8943,  8924,  8905,  8886,  8867,  8849,  8830,
+	8812,  8793,  8775,  8756,  8738,  8720,  8702,  8684,  8666,  8648,  8630,
+	8613,  8595,  8577,  8560,  8542,  8525,  8508,  8490,  8473,  8456,  8439,
+	8422,  8405,  8389,  8372,  8355,  8339,  8322,  8306,  8289,  8273,  8257,
+	8240,  8224,  8208,  8192,
+};
+
+/**
+ * struct vdec_av1_slice_init_vsi - VSI used to initialize instance
+ * @architecture:	architecture type
+ * @reserved:		reserved
+ * @core_vsi:		for core vsi
+ * @cdf_table_addr:	cdf table addr
+ * @cdf_table_size:	cdf table size
+ * @iq_table_addr:	iq table addr
+ * @iq_table_size:	iq table size
+ * @vsi_size:		share vsi structure size
+ */
+struct vdec_av1_slice_init_vsi {
+	u32 architecture;
+	u32 reserved;
+	u64 core_vsi;
+	u64 cdf_table_addr;
+	u32 cdf_table_size;
+	u64 iq_table_addr;
+	u32 iq_table_size;
+	u32 vsi_size;
+};
+
+/**
+ * struct vdec_av1_slice_mem - memory address and size
+ * @buf:		dma_addr padding
+ * @dma_addr:		buffer address
+ * @size:		buffer size
+ * @dma_addr_end:	buffer end address
+ * @padding:		for padding
+ */
+struct vdec_av1_slice_mem {
+	union {
+		u64 buf;
+		dma_addr_t dma_addr;
+	};
+	union {
+		size_t size;
+		dma_addr_t dma_addr_end;
+		u64 padding;
+	};
+};
+
+/**
+ * struct vdec_av1_slice_state - decoding state
+ * @err                   : err type for decode
+ * @full                  : transcoded buffer is full or not
+ * @timeout               : decode timeout or not
+ * @perf                  : performance enable
+ * @crc                   : hw checksum
+ * @out_size              : hw output size
+ */
+struct vdec_av1_slice_state {
+	int err;
+	u32 full;
+	u32 timeout;
+	u32 perf;
+	u32 crc[16];
+	u32 out_size;
+};
+
+/*
+ * enum vdec_av1_slice_resolution_level - resolution level
+ */
+enum vdec_av1_slice_resolution_level {
+	AV1_RES_NONE,
+	AV1_RES_FHD,
+	AV1_RES_4K,
+	AV1_RES_8K,
+};
+
+/*
+ * enum vdec_av1_slice_frame_type - av1 frame type
+ */
+enum vdec_av1_slice_frame_type {
+	AV1_KEY_FRAME = 0,
+	AV1_INTER_FRAME,
+	AV1_INTRA_ONLY_FRAME,
+	AV1_SWITCH_FRAME,
+	AV1_FRAME_TYPES,
+};
+
+/*
+ * enum vdec_av1_slice_reference_mode - reference mode type
+ */
+enum vdec_av1_slice_reference_mode {
+	AV1_SINGLE_REFERENCE = 0,
+	AV1_COMPOUND_REFERENCE,
+	AV1_REFERENCE_MODE_SELECT,
+	AV1_REFERENCE_MODES,
+};
+
+/**
+ * struct vdec_av1_slice_tile_group - info for each tile
+ * @num_tiles:			tile number
+ * @tile_size:			input size for each tile
+ * @tile_start_offset:		tile offset to input buffer
+ */
+struct vdec_av1_slice_tile_group {
+	u32 num_tiles;
+	u32 tile_size[V4L2_AV1_MAX_TILE_COUNT];
+	u32 tile_start_offset[V4L2_AV1_MAX_TILE_COUNT];
+};
+
+/**
+ * struct vdec_av1_slice_scale_factors - scale info for each ref frame
+ * @is_scaled:  frame is scaled or not
+ * @x_scale:    frame width scale coefficient
+ * @y_scale:    frame height scale coefficient
+ * @x_step:     width step for x_scale
+ * @y_step:     height step for y_scale
+ */
+struct vdec_av1_slice_scale_factors {
+	u8 is_scaled;
+	int x_scale;
+	int y_scale;
+	int x_step;
+	int y_step;
+};
+
+/**
+ * struct vdec_av1_slice_frame_refs - ref frame info
+ * @ref_fb_idx:         ref slot index
+ * @ref_map_idx:        ref frame index
+ * @scale_factors:      scale factors for each ref frame
+ */
+struct vdec_av1_slice_frame_refs {
+	int ref_fb_idx;
+	int ref_map_idx;
+	struct vdec_av1_slice_scale_factors scale_factors;
+};
+
+/**
+ * struct vdec_av1_slice_gm - AV1 Global Motion parameters
+ * @wmtype:     The type of global motion transform used
+ * @wmmat:      gm_params
+ * @alpha:      alpha info
+ * @beta:       beta info
+ * @gamma:      gamma info
+ * @delta:      delta info
+ * @invalid:    is invalid or not
+ */
+struct vdec_av1_slice_gm {
+	int wmtype;
+	int wmmat[8];
+	short alpha;
+	short beta;
+	short gamma;
+	short delta;
+	char invalid;
+};
+
+/**
+ * struct vdec_av1_slice_sm - AV1 Skip Mode parameters
+ * @skip_mode_allowed:  Skip Mode is allowed or not
+ * @skip_mode_present:  specified that the skip_mode will be present or not
+ * @skip_mode_frame:    specifies the frames to use for compound prediction
+ */
+struct vdec_av1_slice_sm {
+	u8 skip_mode_allowed;
+	u8 skip_mode_present;
+	int skip_mode_frame[2];
+};
+
+/**
+ * struct vdec_av1_slice_seg - AV1 Segmentation params
+ * @segmentation_enabled:        this frame makes use of the segmentation tool or not
+ * @segmentation_update_map:     segmentation map are updated during the decoding frame
+ * @segmentation_temporal_update:segmentation map are coded relative the existing segmentaion map
+ * @segmentation_update_data:    new parameters are about to be specified for each segment
+ * @feature_data:                specifies the feature data for a segment feature
+ * @feature_enabled_mask:        the corresponding feature value is coded or not.
+ * @segid_preskip:               segment id will be read before the skip syntax element.
+ * @last_active_segid:           the highest numbered segment id that has some enabled feature
+ */
+struct vdec_av1_slice_seg {
+	u8 segmentation_enabled;
+	u8 segmentation_update_map;
+	u8 segmentation_temporal_update;
+	u8 segmentation_update_data;
+	int feature_data[V4L2_AV1_MAX_SEGMENTS][V4L2_AV1_SEG_LVL_MAX];
+	u16 feature_enabled_mask[V4L2_AV1_MAX_SEGMENTS];
+	int segid_preskip;
+	int last_active_segid;
+};
+
+/**
+ * struct vdec_av1_slice_delta_q_lf - AV1 Loop Filter delta parameters
+ * @delta_q_present:    specified whether quantizer index delta values are present
+ * @delta_q_res:        specifies the left shift which should be applied to decoded quantizer index
+ * @delta_lf_present:   specifies whether loop filter delta values are present
+ * @delta_lf_res:       specifies the left shift which should be applied to decoded
+ *                      loop filter delta values
+ * @delta_lf_multi:     specifies that separate loop filter deltas are sent for horizontal
+ *                      luma edges,vertical luma edges,the u edges, and the v edges.
+ */
+struct vdec_av1_slice_delta_q_lf {
+	u8 delta_q_present;
+	u8 delta_q_res;
+	u8 delta_lf_present;
+	u8 delta_lf_res;
+	u8 delta_lf_multi;
+};
+
+/**
+ * struct vdec_av1_slice_quantization - AV1 Quantization params
+ * @base_q_idx:         indicates the base frame qindex. This is used for Y AC
+ *                      coefficients and as the base value for the other quantizers.
+ * @qindex:             qindex
+ * @delta_qydc:         indicates the Y DC quantizer relative to base_q_idx
+ * @delta_qudc:         indicates the U DC quantizer relative to base_q_idx.
+ * @delta_quac:         indicates the U AC quantizer relative to base_q_idx
+ * @delta_qvdc:         indicates the V DC quantizer relative to base_q_idx
+ * @delta_qvac:         indicates the V AC quantizer relative to base_q_idx
+ * @using_qmatrix:      specifies that the quantizer matrix will be used to
+ *                      compute quantizers
+ * @qm_y:               specifies the level in the quantizer matrix that should
+ *                      be used for luma plane decoding
+ * @qm_u:               specifies the level in the quantizer matrix that should
+ *                      be used for chroma U plane decoding.
+ * @qm_v:               specifies the level in the quantizer matrix that should be
+ *                      used for chroma V plane decoding
+ */
+struct vdec_av1_slice_quantization {
+	int base_q_idx;
+	int qindex[V4L2_AV1_MAX_SEGMENTS];
+	int delta_qydc;
+	int delta_qudc;
+	int delta_quac;
+	int delta_qvdc;
+	int delta_qvac;
+	u8 using_qmatrix;
+	u8 qm_y;
+	u8 qm_u;
+	u8 qm_v;
+};
+
+/**
+ * struct vdec_av1_slice_lr - AV1 Loop Restauration parameters
+ * @use_lr:                     whether to use loop restoration
+ * @use_chroma_lr:              whether to use chroma loop restoration
+ * @frame_restoration_type:     specifies the type of restoration used for each plane
+ * @loop_restoration_size:      pecifies the size of loop restoration units in units
+ *                              of samples in the current plane
+ */
+struct vdec_av1_slice_lr {
+	u8 use_lr;
+	u8 use_chroma_lr;
+	u8 frame_restoration_type[V4L2_AV1_NUM_PLANES_MAX];
+	u32 loop_restoration_size[V4L2_AV1_NUM_PLANES_MAX];
+};
+
+/**
+ * struct vdec_av1_slice_loop_filter - AV1 Loop filter parameters
+ * @loop_filter_level:          an array containing loop filter strength values.
+ * @loop_filter_ref_deltas:     contains the adjustment needed for the filter
+ *                              level based on the chosen reference frame
+ * @loop_filter_mode_deltas:    contains the adjustment needed for the filter
+ *                              level based on the chosen mode
+ * @loop_filter_sharpness:      indicates the sharpness level. The loop_filter_level
+ *                              and loop_filter_sharpness together determine when
+ *                              a block edge is filtered, and by how much the
+ *                              filtering can change the sample values
+ * @loop_filter_delta_enabled:  filetr level depends on the mode and reference
+ *                              frame used to predict a block
+ */
+struct vdec_av1_slice_loop_filter {
+	u8 loop_filter_level[4];
+	int loop_filter_ref_deltas[V4L2_AV1_TOTAL_REFS_PER_FRAME];
+	int loop_filter_mode_deltas[2];
+	u8 loop_filter_sharpness;
+	u8 loop_filter_delta_enabled;
+};
+
+/**
+ * struct vdec_av1_slice_cdef - AV1 CDEF parameters
+ * @cdef_damping:       controls the amount of damping in the deringing filter
+ * @cdef_y_strength:    specifies the strength of the primary filter and secondary filter
+ * @cdef_uv_strength:   specifies the strength of the primary filter and secondary filter
+ * @cdef_bits:          specifies the number of bits needed to specify which
+ *                      CDEF filter to apply
+ */
+struct vdec_av1_slice_cdef {
+	u8 cdef_damping;
+	u8 cdef_y_strength[8];
+	u8 cdef_uv_strength[8];
+	u8 cdef_bits;
+};
+
+/**
+ * struct vdec_av1_slice_mfmv - AV1 mfmv parameters
+ * @mfmv_valid_ref:     mfmv_valid_ref
+ * @mfmv_dir:           mfmv_dir
+ * @mfmv_ref_to_cur:    mfmv_ref_to_cur
+ * @mfmv_ref_frame_idx: mfmv_ref_frame_idx
+ * @mfmv_count:         mfmv_count
+ */
+struct vdec_av1_slice_mfmv {
+	u32 mfmv_valid_ref[3];
+	u32 mfmv_dir[3];
+	int mfmv_ref_to_cur[3];
+	int mfmv_ref_frame_idx[3];
+	int mfmv_count;
+};
+
+/**
+ * struct vdec_av1_slice_tile - AV1 Tile info
+ * @tile_cols:                  specifies the number of tiles across the frame
+ * @tile_rows:                  pecifies the number of tiles down the frame
+ * @mi_col_starts:              an array specifying the start column
+ * @mi_row_starts:              an array specifying the start row
+ * @context_update_tile_id:     specifies which tile to use for the CDF update
+ * @uniform_tile_spacing_flag:  tiles are uniformly spaced across the frame
+ *                              or the tile sizes are coded
+ */
+struct vdec_av1_slice_tile {
+	u8 tile_cols;
+	u8 tile_rows;
+	int mi_col_starts[V4L2_AV1_MAX_TILE_COLS + 1];
+	int mi_row_starts[V4L2_AV1_MAX_TILE_ROWS + 1];
+	u8 context_update_tile_id;
+	u8 uniform_tile_spacing_flag;
+};
+
+/**
+ * struct vdec_av1_slice_uncompressed_header - Represents an AV1 Frame Header OBU
+ * @use_ref_frame_mvs:          use_ref_frame_mvs flag
+ * @order_hint:                 specifies OrderHintBits least significant bits of the expected
+ * @gm:                         global motion param
+ * @upscaled_width:             the upscaled width
+ * @frame_width:                frame's width
+ * @frame_height:               frame's height
+ * @reduced_tx_set:             frame is restricted to a reduced subset of the full
+ *                              set of transform types
+ * @tx_mode:                    specifies how the transform size is determined
+ * @uniform_tile_spacing_flag:  tiles are uniformly spaced across the frame
+ *                              or the tile sizes are coded
+ * @interpolation_filter:       specifies the filter selection used for performing inter prediction
+ * @allow_warped_motion:        motion_mode may be present or not
+ * @is_motion_mode_switchable : euqlt to 0 specifies that only the SIMPLE motion mode will be used
+ * @reference_mode :            frame reference mode selected
+ * @allow_high_precision_mv:    specifies that motion vectors are specified to
+ *                              quarter pel precision or to eighth pel precision
+ * @allow_intra_bc:             ubducates that intra block copy may be used in this frame
+ * @force_integer_mv:           specifies motion vectors will always be integers or
+ *                              can contain fractional bits
+ * @allow_screen_content_tools: intra blocks may use palette encoding
+ * @error_resilient_mode:       error resislent mode is enable/disable
+ * @frame_type:                 specifies the AV1 frame type
+ * @primary_ref_frame:          specifies which reference frame contains the CDF values
+ *                              and other state that should be loaded at the start of the frame
+ *                              slots will be updated with the current frame after it is decoded
+ * @disable_frame_end_update_cdf:indicates the end of frame CDF update is disable or enable
+ * @disable_cdf_update:         specified whether the CDF update in the symbol
+ *                              decoding process should be disables
+ * @skip_mode:                  av1 skip mode parameters
+ * @seg:                        av1 segmentaon parameters
+ * @delta_q_lf:                 av1 delta loop fileter
+ * @quant:                      av1 Quantization params
+ * @lr:                         av1 Loop Restauration parameters
+ * @superres_denom:             the denominator for the upscaling ratio
+ * @loop_filter:                av1 Loop filter parameters
+ * @cdef:                       av1 CDEF parameters
+ * @mfmv:                       av1 mfmv parameters
+ * @tile:                       av1 Tile info
+ * @frame_is_intra:             intra frame
+ * @loss_less_array:            loss less array
+ * @coded_loss_less:            coded lsss less
+ * @mi_rows:                    size of mi unit in rows
+ * @mi_cols:                    size of mi unit in cols
+ */
+struct vdec_av1_slice_uncompressed_header {
+	u8 use_ref_frame_mvs;
+	int order_hint;
+	struct vdec_av1_slice_gm gm[V4L2_AV1_TOTAL_REFS_PER_FRAME];
+	u32 upscaled_width;
+	u32 frame_width;
+	u32 frame_height;
+	u8 reduced_tx_set;
+	u8 tx_mode;
+	u8 uniform_tile_spacing_flag;
+	u8 interpolation_filter;
+	u8 allow_warped_motion;
+	u8 is_motion_mode_switchable;
+	u8 reference_mode;
+	u8 allow_high_precision_mv;
+	u8 allow_intra_bc;
+	u8 force_integer_mv;
+	u8 allow_screen_content_tools;
+	u8 error_resilient_mode;
+	u8 frame_type;
+	u8 primary_ref_frame;
+	u8 disable_frame_end_update_cdf;
+	u32 disable_cdf_update;
+	struct vdec_av1_slice_sm skip_mode;
+	struct vdec_av1_slice_seg seg;
+	struct vdec_av1_slice_delta_q_lf delta_q_lf;
+	struct vdec_av1_slice_quantization quant;
+	struct vdec_av1_slice_lr lr;
+	u32 superres_denom;
+	struct vdec_av1_slice_loop_filter loop_filter;
+	struct vdec_av1_slice_cdef cdef;
+	struct vdec_av1_slice_mfmv mfmv;
+	struct vdec_av1_slice_tile tile;
+	u8 frame_is_intra;
+	u8 loss_less_array[V4L2_AV1_MAX_SEGMENTS];
+	u8 coded_loss_less;
+	u32 mi_rows;
+	u32 mi_cols;
+};
+
+/**
+ * struct vdec_av1_slice_seq_header - Represents an AV1 Sequence OBU
+ * @bitdepth:                   the bitdepth to use for the sequence
+ * @enable_superres:            specifies whether the use_superres syntax element may be present
+ * @enable_filter_intra:        specifies the use_filter_intra syntax element may be present
+ * @enable_intra_edge_filter:   whether the intra edge filtering process should be enabled
+ * @enable_interintra_compound: specifies the mode info fo rinter blocks may
+ *                              contain the syntax element interintra
+ * @enable_masked_compound:     specifies the mode info fo rinter blocks may
+ *                              contain the syntax element compound_type
+ * @enable_dual_filter:         the inter prediction filter type may be specified independently
+ * @enable_jnt_comp:            distance weights process may be used for inter prediction
+ * @mono_chrome:                indicates the video does not contain U and V color planes
+ * @enable_order_hint:          tools based on the values of order hints may be used
+ * @order_hint_bits:            the number of bits used for the order_hint field at each frame
+ * @use_128x128_superblock:     indicates superblocks contain 128*128 luma samples
+ * @subsampling_x:              the chroma subsamling format
+ * @subsampling_y:              the chroma subsamling format
+ * @max_frame_width:            the maximum frame width for the frames represented by sequence
+ * @max_frame_height:           the maximum frame height for the frames represented by sequence
+ */
+struct vdec_av1_slice_seq_header {
+	u8 bitdepth;
+	u8 enable_superres;
+	u8 enable_filter_intra;
+	u8 enable_intra_edge_filter;
+	u8 enable_interintra_compound;
+	u8 enable_masked_compound;
+	u8 enable_dual_filter;
+	u8 enable_jnt_comp;
+	u8 mono_chrome;
+	u8 enable_order_hint;
+	u8 order_hint_bits;
+	u8 use_128x128_superblock;
+	u8 subsampling_x;
+	u8 subsampling_y;
+	u32 max_frame_width;
+	u32 max_frame_height;
+};
+
+/**
+ * struct vdec_av1_slice_frame - Represents current Frame info
+ * @uh:                         uncompressed header info
+ * @seq:                        sequence header info
+ * @large_scale_tile:           is large scale mode
+ * @cur_ts:                     current frame timestamp
+ * @prev_fb_idx:                prev slot id
+ * @ref_frame_sign_bias:        arrays for ref_frame sign bias
+ * @order_hints:                arrays for ref_frame order hint
+ * @ref_frame_valid:            arrays for valid ref_frame
+ * @ref_frame_map:              map to slot frame info
+ * @frame_refs:                 ref_frame info
+ */
+struct vdec_av1_slice_frame {
+	struct vdec_av1_slice_uncompressed_header uh;
+	struct vdec_av1_slice_seq_header seq;
+	u8 large_scale_tile;
+	u64 cur_ts;
+	int prev_fb_idx;
+	u8 ref_frame_sign_bias[V4L2_AV1_TOTAL_REFS_PER_FRAME];
+	u32 order_hints[V4L2_AV1_REFS_PER_FRAME];
+	u32 ref_frame_valid[V4L2_AV1_REFS_PER_FRAME];
+	int ref_frame_map[V4L2_AV1_TOTAL_REFS_PER_FRAME];
+	struct vdec_av1_slice_frame_refs frame_refs[V4L2_AV1_REFS_PER_FRAME];
+};
+
+/**
+ * struct vdec_av1_slice_work_buffer - work buffer for lat
+ * @mv_addr:    mv buffer memory info
+ * @cdf_addr:   cdf buffer memory info
+ * @segid_addr: segid buffer memory info
+ */
+struct vdec_av1_slice_work_buffer {
+	struct vdec_av1_slice_mem mv_addr;
+	struct vdec_av1_slice_mem cdf_addr;
+	struct vdec_av1_slice_mem segid_addr;
+};
+
+/**
+ * struct vdec_av1_slice_frame_info - frame info for each slot
+ * @frame_type:         frame type
+ * @frame_is_intra:     is intra frame
+ * @order_hint:         order hint
+ * @order_hints:        referece frame order hint
+ * @upscaled_width:     upscale width
+ * @pic_pitch:          buffer pitch
+ * @frame_width:        frane width
+ * @frame_height:       frame height
+ * @mi_rows:            rows in mode info
+ * @mi_cols:            cols in mode info
+ * @ref_count:          mark to reference frame counts
+ */
+struct vdec_av1_slice_frame_info {
+	u8 frame_type;
+	u8 frame_is_intra;
+	int order_hint;
+	u32 order_hints[V4L2_AV1_REFS_PER_FRAME];
+	u32 upscaled_width;
+	u32 pic_pitch;
+	u32 frame_width;
+	u32 frame_height;
+	u32 mi_rows;
+	u32 mi_cols;
+	int ref_count;
+};
+
+/**
+ * struct vdec_av1_slice_slot - slot info that needs to be saved in the global instance
+ * @frame_info: frame info for each slot
+ * @timestamp:  time stamp info
+ */
+struct vdec_av1_slice_slot {
+	struct vdec_av1_slice_frame_info frame_info[AV1_MAX_FRAME_BUF_COUNT];
+	u64 timestamp[AV1_MAX_FRAME_BUF_COUNT];
+};
+
+/**
+ * struct vdec_av1_slice_fb - frame buffer for decoding
+ * @y:  current y buffer address info
+ * @c:  current c buffer address info
+ */
+struct vdec_av1_slice_fb {
+	struct vdec_av1_slice_mem y;
+	struct vdec_av1_slice_mem c;
+};
+
+/**
+ * struct vdec_av1_slice_vsi - exchange frame information between Main CPU and MicroP
+ * @bs:			input buffer info
+ * @work_buffer:	working buffe for hw
+ * @cdf_table:		cdf_table buffer
+ * @cdf_tmp:		cdf temp buffer
+ * @rd_mv:		mv buffer for lat output , core input
+ * @ube:		ube buffer
+ * @trans:		transcoded buffer
+ * @err_map:		err map buffer
+ * @row_info:		row info buffer
+ * @fb:			current y/c buffer
+ * @ref:		ref y/c buffer
+ * @iq_table:		iq table buffer
+ * @tile:		tile buffer
+ * @slots:		slots info for each frame
+ * @slot_id:		current frame slot id
+ * @frame:		current frame info
+ * @state:		status after decode done
+ * @cur_lst_tile_id:	tile id for large scale
+ */
+struct vdec_av1_slice_vsi {
+	/* lat */
+	struct vdec_av1_slice_mem bs;
+	struct vdec_av1_slice_work_buffer work_buffer[AV1_MAX_FRAME_BUF_COUNT];
+	struct vdec_av1_slice_mem cdf_table;
+	struct vdec_av1_slice_mem cdf_tmp;
+	/* LAT stage's output, Core stage's input */
+	struct vdec_av1_slice_mem rd_mv;
+	struct vdec_av1_slice_mem ube;
+	struct vdec_av1_slice_mem trans;
+	struct vdec_av1_slice_mem err_map;
+	struct vdec_av1_slice_mem row_info;
+	/* core */
+	struct vdec_av1_slice_fb fb;
+	struct vdec_av1_slice_fb ref[V4L2_AV1_REFS_PER_FRAME];
+	struct vdec_av1_slice_mem iq_table;
+	/* lat and core share*/
+	struct vdec_av1_slice_mem tile;
+	struct vdec_av1_slice_slot slots;
+	s8 slot_id;
+	struct vdec_av1_slice_frame frame;
+	struct vdec_av1_slice_state state;
+	u32 cur_lst_tile_id;
+};
+
+/**
+ * struct vdec_av1_slice_pfc - per-frame context that contains a local vsi.
+ *                             pass it from lat to core
+ * @vsi:        local vsi. copy to/from remote vsi before/after decoding
+ * @ref_idx:    reference buffer timestamp
+ * @seq:        picture sequence
+ */
+struct vdec_av1_slice_pfc {
+	struct vdec_av1_slice_vsi vsi;
+	u64 ref_idx[V4L2_AV1_REFS_PER_FRAME];
+	int seq;
+};
+
+/**
+ * struct vdec_av1_slice_instance - represent one av1 instance
+ * @ctx:                pointer to codec's context
+ * @vpu:                VPU instance
+ * @iq_table:           iq table buffer
+ * @cdf_table:          cdf table buffer
+ * @mv:                 mv working buffer
+ * @cdf:                cdf working buffer
+ * @seg:                segmentation working buffer
+ * @cdf_temp:           cdf temp buffer
+ * @tile:               tile buffer
+ * @slots:              slots info
+ * @tile_group:         tile_group entry
+ * @level:              level of current resolution
+ * @width:              width of last picture
+ * @height:             height of last picture
+ * @frame_type:         frame_type of last picture
+ * @irq_enabled:        irq to Main CPU or MicroP
+ * @inneracing_mode:    is inneracing mode
+ * @init_vsi:           vsi used for initialized AV1 instance
+ * @vsi:                vsi used for decoding/flush ...
+ * @core_vsi:           vsi used for Core stage
+ * @seq:                global picture sequence
+ */
+struct vdec_av1_slice_instance {
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vpu_inst vpu;
+
+	struct mtk_vcodec_mem iq_table;
+	struct mtk_vcodec_mem cdf_table;
+
+	struct mtk_vcodec_mem mv[AV1_MAX_FRAME_BUF_COUNT];
+	struct mtk_vcodec_mem cdf[AV1_MAX_FRAME_BUF_COUNT];
+	struct mtk_vcodec_mem seg[AV1_MAX_FRAME_BUF_COUNT];
+	struct mtk_vcodec_mem cdf_temp;
+	struct mtk_vcodec_mem tile;
+	struct vdec_av1_slice_slot slots;
+	struct vdec_av1_slice_tile_group tile_group;
+
+	/* for resolution change and get_pic_info */
+	enum vdec_av1_slice_resolution_level level;
+	u32 width;
+	u32 height;
+
+	u32 frame_type;
+	u32 irq_enabled;
+	u32 inneracing_mode;
+
+	/* MicroP vsi */
+	union {
+		struct vdec_av1_slice_init_vsi *init_vsi;
+		struct vdec_av1_slice_vsi *vsi;
+	};
+	struct vdec_av1_slice_vsi *core_vsi;
+	int seq;
+};
+
+static int vdec_av1_slice_core_decode(struct vdec_lat_buf *lat_buf);
+
+static inline int vdec_av1_slice_get_msb(u32 n)
+{
+	if (n == 0)
+		return 0;
+	return 31 ^ __builtin_clz(n);
+}
+
+static inline bool vdec_av1_slice_need_scale(u32 ref_width, u32 ref_height,
+					     u32 this_width, u32 this_height)
+{
+	return ((this_width << 1) >= ref_width) &&
+		((this_height << 1) >= ref_height) &&
+		(this_width <= (ref_width << 4)) &&
+		(this_height <= (ref_height << 4));
+}
+
+static void *vdec_av1_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id)
+{
+	struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id);
+
+	if (!ctrl)
+		return ERR_PTR(-EINVAL);
+
+	return ctrl->p_cur.p;
+}
+
+static int vdec_av1_slice_init_cdf_table(struct vdec_av1_slice_instance *instance)
+{
+	u8 *remote_cdf_table;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_av1_slice_init_vsi *vsi;
+	int ret;
+
+	ctx = instance->ctx;
+	vsi = instance->vpu.vsi;
+	remote_cdf_table = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler,
+						     (u32)vsi->cdf_table_addr);
+	if (IS_ERR(remote_cdf_table)) {
+		mtk_vdec_err(ctx, "failed to map cdf table\n");
+		return PTR_ERR(remote_cdf_table);
+	}
+
+	mtk_vdec_debug(ctx, "map cdf table to 0x%p\n", remote_cdf_table);
+
+	if (instance->cdf_table.va)
+		mtk_vcodec_mem_free(ctx, &instance->cdf_table);
+	instance->cdf_table.size = vsi->cdf_table_size;
+
+	ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf_table);
+	if (ret)
+		return ret;
+
+	memcpy(instance->cdf_table.va, remote_cdf_table, vsi->cdf_table_size);
+
+	return 0;
+}
+
+static int vdec_av1_slice_init_iq_table(struct vdec_av1_slice_instance *instance)
+{
+	u8 *remote_iq_table;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_av1_slice_init_vsi *vsi;
+	int ret;
+
+	ctx = instance->ctx;
+	vsi = instance->vpu.vsi;
+	remote_iq_table = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler,
+						    (u32)vsi->iq_table_addr);
+	if (IS_ERR(remote_iq_table)) {
+		mtk_vdec_err(ctx, "failed to map iq table\n");
+		return PTR_ERR(remote_iq_table);
+	}
+
+	mtk_vdec_debug(ctx, "map iq table to 0x%p\n", remote_iq_table);
+
+	if (instance->iq_table.va)
+		mtk_vcodec_mem_free(ctx, &instance->iq_table);
+	instance->iq_table.size = vsi->iq_table_size;
+
+	ret = mtk_vcodec_mem_alloc(ctx, &instance->iq_table);
+	if (ret)
+		return ret;
+
+	memcpy(instance->iq_table.va, remote_iq_table, vsi->iq_table_size);
+
+	return 0;
+}
+
+static int vdec_av1_slice_get_new_slot(struct vdec_av1_slice_vsi *vsi)
+{
+	struct vdec_av1_slice_slot *slots = &vsi->slots;
+	int new_slot_idx = AV1_INVALID_IDX;
+	int i;
+
+	for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) {
+		if (slots->frame_info[i].ref_count == 0) {
+			new_slot_idx = i;
+			break;
+		}
+	}
+
+	if (new_slot_idx != AV1_INVALID_IDX) {
+		slots->frame_info[new_slot_idx].ref_count++;
+		slots->timestamp[new_slot_idx] = vsi->frame.cur_ts;
+	}
+
+	return new_slot_idx;
+}
+
+static inline void vdec_av1_slice_clear_fb(struct vdec_av1_slice_frame_info *frame_info)
+{
+	memset((void *)frame_info, 0, sizeof(struct vdec_av1_slice_frame_info));
+}
+
+static void vdec_av1_slice_decrease_ref_count(struct vdec_av1_slice_slot *slots, int fb_idx)
+{
+	struct vdec_av1_slice_frame_info *frame_info = slots->frame_info;
+
+	frame_info[fb_idx].ref_count--;
+	if (frame_info[fb_idx].ref_count < 0) {
+		frame_info[fb_idx].ref_count = 0;
+		pr_err(MTK_DBG_V4L2_STR "av1_error: %s() fb_idx %d decrease ref_count error\n",
+		       __func__, fb_idx);
+	}
+
+	vdec_av1_slice_clear_fb(&frame_info[fb_idx]);
+}
+
+static void vdec_av1_slice_cleanup_slots(struct vdec_av1_slice_slot *slots,
+					 struct vdec_av1_slice_frame *frame,
+					 struct v4l2_ctrl_av1_frame *ctrl_fh)
+{
+	int slot_id, ref_id;
+
+	for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++)
+		frame->ref_frame_map[ref_id] = AV1_INVALID_IDX;
+
+	for (slot_id = 0; slot_id < AV1_MAX_FRAME_BUF_COUNT; slot_id++) {
+		u64 timestamp = slots->timestamp[slot_id];
+		bool ref_used = false;
+
+		/* ignored unused slots */
+		if (slots->frame_info[slot_id].ref_count == 0)
+			continue;
+
+		for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++) {
+			if (ctrl_fh->reference_frame_ts[ref_id] == timestamp) {
+				frame->ref_frame_map[ref_id] = slot_id;
+				ref_used = true;
+			}
+		}
+
+		if (!ref_used)
+			vdec_av1_slice_decrease_ref_count(slots, slot_id);
+	}
+}
+
+static void vdec_av1_slice_setup_slot(struct vdec_av1_slice_instance *instance,
+				      struct vdec_av1_slice_vsi *vsi,
+				      struct v4l2_ctrl_av1_frame *ctrl_fh)
+{
+	struct vdec_av1_slice_frame_info *cur_frame_info;
+	struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh;
+	int ref_id;
+
+	memcpy(&vsi->slots, &instance->slots, sizeof(instance->slots));
+	vdec_av1_slice_cleanup_slots(&vsi->slots, &vsi->frame, ctrl_fh);
+	vsi->slot_id = vdec_av1_slice_get_new_slot(vsi);
+
+	if (vsi->slot_id == AV1_INVALID_IDX) {
+		mtk_v4l2_vdec_err(instance->ctx, "warning:av1 get invalid index slot\n");
+		vsi->slot_id = 0;
+	}
+	cur_frame_info = &vsi->slots.frame_info[vsi->slot_id];
+	cur_frame_info->frame_type = uh->frame_type;
+	cur_frame_info->frame_is_intra = ((uh->frame_type == AV1_INTRA_ONLY_FRAME) ||
+					  (uh->frame_type == AV1_KEY_FRAME));
+	cur_frame_info->order_hint = uh->order_hint;
+	cur_frame_info->upscaled_width = uh->upscaled_width;
+	cur_frame_info->pic_pitch = 0;
+	cur_frame_info->frame_width = uh->frame_width;
+	cur_frame_info->frame_height = uh->frame_height;
+	cur_frame_info->mi_cols = ((uh->frame_width + 7) >> 3) << 1;
+	cur_frame_info->mi_rows = ((uh->frame_height + 7) >> 3) << 1;
+
+	/* ensure current frame is properly mapped if referenced */
+	for (ref_id = 0; ref_id < V4L2_AV1_TOTAL_REFS_PER_FRAME; ref_id++) {
+		u64 timestamp = vsi->slots.timestamp[vsi->slot_id];
+
+		if (ctrl_fh->reference_frame_ts[ref_id] == timestamp)
+			vsi->frame.ref_frame_map[ref_id] = vsi->slot_id;
+	}
+}
+
+static int vdec_av1_slice_alloc_working_buffer(struct vdec_av1_slice_instance *instance,
+					       struct vdec_av1_slice_vsi *vsi)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	enum vdec_av1_slice_resolution_level level;
+	u32 max_sb_w, max_sb_h, max_w, max_h, w, h;
+	int i, ret;
+
+	w = vsi->frame.uh.frame_width;
+	h = vsi->frame.uh.frame_height;
+
+	if (w > VCODEC_DEC_4K_CODED_WIDTH || h > VCODEC_DEC_4K_CODED_HEIGHT)
+		/* 8K */
+		return -EINVAL;
+
+	if (w > MTK_VDEC_MAX_W || h > MTK_VDEC_MAX_H) {
+		/* 4K */
+		level = AV1_RES_4K;
+		max_w = VCODEC_DEC_4K_CODED_WIDTH;
+		max_h = VCODEC_DEC_4K_CODED_HEIGHT;
+	} else {
+		/* FHD */
+		level = AV1_RES_FHD;
+		max_w = MTK_VDEC_MAX_W;
+		max_h = MTK_VDEC_MAX_H;
+	}
+
+	if (level == instance->level)
+		return 0;
+
+	mtk_vdec_debug(ctx, "resolution level changed from %u to %u, %ux%u",
+		       instance->level, level, w, h);
+
+	max_sb_w = DIV_ROUND_UP(max_w, 128);
+	max_sb_h = DIV_ROUND_UP(max_h, 128);
+
+	for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) {
+		if (instance->mv[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->mv[i]);
+		instance->mv[i].size = max_sb_w * max_sb_h * SZ_1K;
+		ret = mtk_vcodec_mem_alloc(ctx, &instance->mv[i]);
+		if (ret)
+			goto err;
+
+		if (instance->seg[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->seg[i]);
+		instance->seg[i].size = max_sb_w * max_sb_h * 512;
+		ret = mtk_vcodec_mem_alloc(ctx, &instance->seg[i]);
+		if (ret)
+			goto err;
+
+		if (instance->cdf[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->cdf[i]);
+		instance->cdf[i].size = AV1_CDF_TABLE_BUFFER_SIZE;
+		ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf[i]);
+		if (ret)
+			goto err;
+	}
+
+	if (!instance->cdf_temp.va) {
+		instance->cdf_temp.size = (SZ_1K * 16 * 100);
+		ret = mtk_vcodec_mem_alloc(ctx, &instance->cdf_temp);
+		if (ret)
+			goto err;
+		vsi->cdf_tmp.buf = instance->cdf_temp.dma_addr;
+		vsi->cdf_tmp.size = instance->cdf_temp.size;
+	}
+
+	if (instance->tile.va)
+		mtk_vcodec_mem_free(ctx, &instance->tile);
+
+	instance->tile.size = AV1_TILE_BUF_SIZE * V4L2_AV1_MAX_TILE_COUNT;
+	ret = mtk_vcodec_mem_alloc(ctx, &instance->tile);
+	if (ret)
+		goto err;
+
+	instance->level = level;
+	return 0;
+
+err:
+	instance->level = AV1_RES_NONE;
+	return ret;
+}
+
+static void vdec_av1_slice_free_working_buffer(struct vdec_av1_slice_instance *instance)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(instance->mv); i++)
+		mtk_vcodec_mem_free(ctx, &instance->mv[i]);
+
+	for (i = 0; i < ARRAY_SIZE(instance->seg); i++)
+		mtk_vcodec_mem_free(ctx, &instance->seg[i]);
+
+	for (i = 0; i < ARRAY_SIZE(instance->cdf); i++)
+		mtk_vcodec_mem_free(ctx, &instance->cdf[i]);
+
+	mtk_vcodec_mem_free(ctx, &instance->tile);
+	mtk_vcodec_mem_free(ctx, &instance->cdf_temp);
+	mtk_vcodec_mem_free(ctx, &instance->cdf_table);
+	mtk_vcodec_mem_free(ctx, &instance->iq_table);
+
+	instance->level = AV1_RES_NONE;
+}
+
+static inline void vdec_av1_slice_vsi_from_remote(struct vdec_av1_slice_vsi *vsi,
+						  struct vdec_av1_slice_vsi *remote_vsi)
+{
+	memcpy(&vsi->trans, &remote_vsi->trans, sizeof(vsi->trans));
+	memcpy(&vsi->state, &remote_vsi->state, sizeof(vsi->state));
+}
+
+static inline void vdec_av1_slice_vsi_to_remote(struct vdec_av1_slice_vsi *vsi,
+						struct vdec_av1_slice_vsi *remote_vsi)
+{
+	memcpy(remote_vsi, vsi, sizeof(*vsi));
+}
+
+static int vdec_av1_slice_setup_lat_from_src_buf(struct vdec_av1_slice_instance *instance,
+						 struct vdec_av1_slice_vsi *vsi,
+						 struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_v4l2_buffer *src;
+	struct vb2_v4l2_buffer *dst;
+
+	src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx);
+	if (!src)
+		return -EINVAL;
+
+	lat_buf->src_buf_req = src->vb2_buf.req_obj.req;
+	dst = &lat_buf->ts_info;
+	v4l2_m2m_buf_copy_metadata(src, dst, true);
+	vsi->frame.cur_ts = dst->vb2_buf.timestamp;
+
+	return 0;
+}
+
+static short vdec_av1_slice_resolve_divisor_32(u32 D, short *shift)
+{
+	int f;
+	int e;
+
+	*shift = vdec_av1_slice_get_msb(D);
+	/* e is obtained from D after resetting the most significant 1 bit. */
+	e = D - ((u32)1 << *shift);
+	/* Get the most significant DIV_LUT_BITS (8) bits of e into f */
+	if (*shift > DIV_LUT_BITS)
+		f = AV1_DIV_ROUND_UP_POW2(e, *shift - DIV_LUT_BITS);
+	else
+		f = e << (DIV_LUT_BITS - *shift);
+	if (f > DIV_LUT_NUM)
+		return -1;
+	*shift += DIV_LUT_PREC_BITS;
+	/* Use f as lookup into the precomputed table of multipliers */
+	return div_lut[f];
+}
+
+static void vdec_av1_slice_get_shear_params(struct vdec_av1_slice_gm *gm_params)
+{
+	const int *mat = gm_params->wmmat;
+	short shift;
+	short y;
+	long long gv, dv;
+
+	if (gm_params->wmmat[2] <= 0)
+		return;
+
+	gm_params->alpha = clamp_val(mat[2] - (1 << WARPEDMODEL_PREC_BITS), S16_MIN, S16_MAX);
+	gm_params->beta = clamp_val(mat[3], S16_MIN, S16_MAX);
+
+	y = vdec_av1_slice_resolve_divisor_32(abs(mat[2]), &shift) * (mat[2] < 0 ? -1 : 1);
+
+	gv = ((long long)mat[4] * (1 << WARPEDMODEL_PREC_BITS)) * y;
+	gm_params->gamma = clamp_val((int)AV1_DIV_ROUND_UP_POW2_SIGNED(gv, shift),
+				     S16_MIN, S16_MAX);
+
+	dv = ((long long)mat[3] * mat[4]) * y;
+	gm_params->delta = clamp_val(mat[5] - (int)AV1_DIV_ROUND_UP_POW2_SIGNED(dv, shift) -
+				     (1 << WARPEDMODEL_PREC_BITS), S16_MIN, S16_MAX);
+
+	gm_params->alpha = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->alpha, WARP_PARAM_REDUCE_BITS) *
+							(1 << WARP_PARAM_REDUCE_BITS);
+	gm_params->beta = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->beta, WARP_PARAM_REDUCE_BITS) *
+						       (1 << WARP_PARAM_REDUCE_BITS);
+	gm_params->gamma = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->gamma, WARP_PARAM_REDUCE_BITS) *
+							(1 << WARP_PARAM_REDUCE_BITS);
+	gm_params->delta = AV1_DIV_ROUND_UP_POW2_SIGNED(gm_params->delta, WARP_PARAM_REDUCE_BITS) *
+							(1 << WARP_PARAM_REDUCE_BITS);
+}
+
+static void vdec_av1_slice_setup_gm(struct vdec_av1_slice_gm *gm,
+				    struct v4l2_av1_global_motion *ctrl_gm)
+{
+	u32 i, j;
+
+	for (i = 0; i < V4L2_AV1_TOTAL_REFS_PER_FRAME; i++) {
+		gm[i].wmtype = ctrl_gm->type[i];
+		for (j = 0; j < 6; j++)
+			gm[i].wmmat[j] = ctrl_gm->params[i][j];
+
+		gm[i].invalid = !!(ctrl_gm->invalid & BIT(i));
+		gm[i].alpha = 0;
+		gm[i].beta = 0;
+		gm[i].gamma = 0;
+		gm[i].delta = 0;
+		if (gm[i].wmtype <= V4L2_AV1_WARP_MODEL_AFFINE)
+			vdec_av1_slice_get_shear_params(&gm[i]);
+	}
+}
+
+static void vdec_av1_slice_setup_seg(struct vdec_av1_slice_seg *seg,
+				     struct v4l2_av1_segmentation *ctrl_seg)
+{
+	u32 i, j;
+
+	seg->segmentation_enabled = SEGMENTATION_FLAG(ctrl_seg, ENABLED);
+	seg->segmentation_update_map = SEGMENTATION_FLAG(ctrl_seg, UPDATE_MAP);
+	seg->segmentation_temporal_update = SEGMENTATION_FLAG(ctrl_seg, TEMPORAL_UPDATE);
+	seg->segmentation_update_data = SEGMENTATION_FLAG(ctrl_seg, UPDATE_DATA);
+	seg->segid_preskip = SEGMENTATION_FLAG(ctrl_seg, SEG_ID_PRE_SKIP);
+	seg->last_active_segid = ctrl_seg->last_active_seg_id;
+
+	for (i = 0; i < V4L2_AV1_MAX_SEGMENTS; i++) {
+		seg->feature_enabled_mask[i] = ctrl_seg->feature_enabled[i];
+		for (j = 0; j < V4L2_AV1_SEG_LVL_MAX; j++)
+			seg->feature_data[i][j] = ctrl_seg->feature_data[i][j];
+	}
+}
+
+static void vdec_av1_slice_setup_quant(struct vdec_av1_slice_quantization *quant,
+				       struct v4l2_av1_quantization *ctrl_quant)
+{
+	quant->base_q_idx = ctrl_quant->base_q_idx;
+	quant->delta_qydc = ctrl_quant->delta_q_y_dc;
+	quant->delta_qudc = ctrl_quant->delta_q_u_dc;
+	quant->delta_quac = ctrl_quant->delta_q_u_ac;
+	quant->delta_qvdc = ctrl_quant->delta_q_v_dc;
+	quant->delta_qvac = ctrl_quant->delta_q_v_ac;
+	quant->qm_y = ctrl_quant->qm_y;
+	quant->qm_u = ctrl_quant->qm_u;
+	quant->qm_v = ctrl_quant->qm_v;
+	quant->using_qmatrix = QUANT_FLAG(ctrl_quant, USING_QMATRIX);
+}
+
+static int vdec_av1_slice_get_qindex(struct vdec_av1_slice_uncompressed_header *uh,
+				     int segmentation_id)
+{
+	struct vdec_av1_slice_seg *seg = &uh->seg;
+	struct vdec_av1_slice_quantization *quant = &uh->quant;
+	int data = 0, qindex = 0;
+
+	if (seg->segmentation_enabled &&
+	    (seg->feature_enabled_mask[segmentation_id] & BIT(SEG_LVL_ALT_Q))) {
+		data = seg->feature_data[segmentation_id][SEG_LVL_ALT_Q];
+		qindex = quant->base_q_idx + data;
+		return clamp_val(qindex, 0, MAXQ);
+	}
+
+	return quant->base_q_idx;
+}
+
+static void vdec_av1_slice_setup_lr(struct vdec_av1_slice_lr *lr,
+				    struct v4l2_av1_loop_restoration  *ctrl_lr)
+{
+	int i;
+
+	lr->use_lr = 0;
+	lr->use_chroma_lr = 0;
+	for (i = 0; i < V4L2_AV1_NUM_PLANES_MAX; i++) {
+		lr->frame_restoration_type[i] = ctrl_lr->frame_restoration_type[i];
+		lr->loop_restoration_size[i] = ctrl_lr->loop_restoration_size[i];
+		if (lr->frame_restoration_type[i]) {
+			lr->use_lr = 1;
+			if (i > 0)
+				lr->use_chroma_lr = 1;
+		}
+	}
+}
+
+static void vdec_av1_slice_setup_lf(struct vdec_av1_slice_loop_filter *lf,
+				    struct v4l2_av1_loop_filter *ctrl_lf)
+{
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(lf->loop_filter_level); i++)
+		lf->loop_filter_level[i] = ctrl_lf->level[i];
+
+	for (i = 0; i < V4L2_AV1_TOTAL_REFS_PER_FRAME; i++)
+		lf->loop_filter_ref_deltas[i] = ctrl_lf->ref_deltas[i];
+
+	for (i = 0; i < ARRAY_SIZE(lf->loop_filter_mode_deltas); i++)
+		lf->loop_filter_mode_deltas[i] = ctrl_lf->mode_deltas[i];
+
+	lf->loop_filter_sharpness = ctrl_lf->sharpness;
+	lf->loop_filter_delta_enabled =
+		   BIT_FLAG(ctrl_lf, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_ENABLED);
+}
+
+static void vdec_av1_slice_setup_cdef(struct vdec_av1_slice_cdef *cdef,
+				      struct v4l2_av1_cdef *ctrl_cdef)
+{
+	int i;
+
+	cdef->cdef_damping = ctrl_cdef->damping_minus_3 + 3;
+	cdef->cdef_bits = ctrl_cdef->bits;
+
+	for (i = 0; i < V4L2_AV1_CDEF_MAX; i++) {
+		if (ctrl_cdef->y_sec_strength[i] == 4)
+			ctrl_cdef->y_sec_strength[i] -= 1;
+
+		if (ctrl_cdef->uv_sec_strength[i] == 4)
+			ctrl_cdef->uv_sec_strength[i] -= 1;
+
+		cdef->cdef_y_strength[i] =
+			ctrl_cdef->y_pri_strength[i] << SECONDARY_FILTER_STRENGTH_NUM_BITS |
+			ctrl_cdef->y_sec_strength[i];
+		cdef->cdef_uv_strength[i] =
+			ctrl_cdef->uv_pri_strength[i] << SECONDARY_FILTER_STRENGTH_NUM_BITS |
+			ctrl_cdef->uv_sec_strength[i];
+	}
+}
+
+static void vdec_av1_slice_setup_seq(struct vdec_av1_slice_seq_header *seq,
+				     struct v4l2_ctrl_av1_sequence *ctrl_seq)
+{
+	seq->bitdepth = ctrl_seq->bit_depth;
+	seq->max_frame_width = ctrl_seq->max_frame_width_minus_1 + 1;
+	seq->max_frame_height = ctrl_seq->max_frame_height_minus_1 + 1;
+	seq->enable_superres = SEQUENCE_FLAG(ctrl_seq, ENABLE_SUPERRES);
+	seq->enable_filter_intra = SEQUENCE_FLAG(ctrl_seq, ENABLE_FILTER_INTRA);
+	seq->enable_intra_edge_filter = SEQUENCE_FLAG(ctrl_seq, ENABLE_INTRA_EDGE_FILTER);
+	seq->enable_interintra_compound = SEQUENCE_FLAG(ctrl_seq, ENABLE_INTERINTRA_COMPOUND);
+	seq->enable_masked_compound = SEQUENCE_FLAG(ctrl_seq, ENABLE_MASKED_COMPOUND);
+	seq->enable_dual_filter = SEQUENCE_FLAG(ctrl_seq, ENABLE_DUAL_FILTER);
+	seq->enable_jnt_comp = SEQUENCE_FLAG(ctrl_seq, ENABLE_JNT_COMP);
+	seq->mono_chrome = SEQUENCE_FLAG(ctrl_seq, MONO_CHROME);
+	seq->enable_order_hint = SEQUENCE_FLAG(ctrl_seq, ENABLE_ORDER_HINT);
+	seq->order_hint_bits = ctrl_seq->order_hint_bits;
+	seq->use_128x128_superblock = SEQUENCE_FLAG(ctrl_seq, USE_128X128_SUPERBLOCK);
+	seq->subsampling_x = SEQUENCE_FLAG(ctrl_seq, SUBSAMPLING_X);
+	seq->subsampling_y = SEQUENCE_FLAG(ctrl_seq, SUBSAMPLING_Y);
+}
+
+static void vdec_av1_slice_setup_tile(struct vdec_av1_slice_frame *frame,
+				      struct v4l2_av1_tile_info *ctrl_tile)
+{
+	struct vdec_av1_slice_seq_header *seq = &frame->seq;
+	struct vdec_av1_slice_tile *tile = &frame->uh.tile;
+	u32 mib_size_log2 = seq->use_128x128_superblock ? 5 : 4;
+	int i;
+
+	tile->tile_cols = ctrl_tile->tile_cols;
+	tile->tile_rows = ctrl_tile->tile_rows;
+	tile->context_update_tile_id = ctrl_tile->context_update_tile_id;
+	tile->uniform_tile_spacing_flag =
+		BIT_FLAG(ctrl_tile, V4L2_AV1_TILE_INFO_FLAG_UNIFORM_TILE_SPACING);
+
+	for (i = 0; i < tile->tile_cols + 1; i++)
+		tile->mi_col_starts[i] =
+			ALIGN(ctrl_tile->mi_col_starts[i], BIT(mib_size_log2)) >> mib_size_log2;
+
+	for (i = 0; i < tile->tile_rows + 1; i++)
+		tile->mi_row_starts[i] =
+			ALIGN(ctrl_tile->mi_row_starts[i], BIT(mib_size_log2)) >> mib_size_log2;
+}
+
+static void vdec_av1_slice_setup_uh(struct vdec_av1_slice_instance *instance,
+				    struct vdec_av1_slice_frame *frame,
+				    struct v4l2_ctrl_av1_frame *ctrl_fh)
+{
+	struct vdec_av1_slice_uncompressed_header *uh = &frame->uh;
+	int i;
+
+	uh->use_ref_frame_mvs = FH_FLAG(ctrl_fh, USE_REF_FRAME_MVS);
+	uh->order_hint = ctrl_fh->order_hint;
+	vdec_av1_slice_setup_gm(uh->gm, &ctrl_fh->global_motion);
+	uh->upscaled_width = ctrl_fh->upscaled_width;
+	uh->frame_width = ctrl_fh->frame_width_minus_1 + 1;
+	uh->frame_height = ctrl_fh->frame_height_minus_1 + 1;
+	uh->mi_cols = ((uh->frame_width + 7) >> 3) << 1;
+	uh->mi_rows = ((uh->frame_height + 7) >> 3) << 1;
+	uh->reduced_tx_set = FH_FLAG(ctrl_fh, REDUCED_TX_SET);
+	uh->tx_mode = ctrl_fh->tx_mode;
+	uh->uniform_tile_spacing_flag =
+		BIT_FLAG(&ctrl_fh->tile_info, V4L2_AV1_TILE_INFO_FLAG_UNIFORM_TILE_SPACING);
+	uh->interpolation_filter = ctrl_fh->interpolation_filter;
+	uh->allow_warped_motion = FH_FLAG(ctrl_fh, ALLOW_WARPED_MOTION);
+	uh->is_motion_mode_switchable = FH_FLAG(ctrl_fh, IS_MOTION_MODE_SWITCHABLE);
+	uh->frame_type = ctrl_fh->frame_type;
+	uh->frame_is_intra = (uh->frame_type == V4L2_AV1_INTRA_ONLY_FRAME ||
+			      uh->frame_type == V4L2_AV1_KEY_FRAME);
+
+	if (!uh->frame_is_intra && FH_FLAG(ctrl_fh, REFERENCE_SELECT))
+		uh->reference_mode = AV1_REFERENCE_MODE_SELECT;
+	else
+		uh->reference_mode = AV1_SINGLE_REFERENCE;
+
+	uh->allow_high_precision_mv = FH_FLAG(ctrl_fh, ALLOW_HIGH_PRECISION_MV);
+	uh->allow_intra_bc = FH_FLAG(ctrl_fh, ALLOW_INTRABC);
+	uh->force_integer_mv = FH_FLAG(ctrl_fh, FORCE_INTEGER_MV);
+	uh->allow_screen_content_tools = FH_FLAG(ctrl_fh, ALLOW_SCREEN_CONTENT_TOOLS);
+	uh->error_resilient_mode = FH_FLAG(ctrl_fh, ERROR_RESILIENT_MODE);
+	uh->primary_ref_frame = ctrl_fh->primary_ref_frame;
+	uh->disable_frame_end_update_cdf =
+			FH_FLAG(ctrl_fh, DISABLE_FRAME_END_UPDATE_CDF);
+	uh->disable_cdf_update = FH_FLAG(ctrl_fh, DISABLE_CDF_UPDATE);
+	uh->skip_mode.skip_mode_present = FH_FLAG(ctrl_fh, SKIP_MODE_PRESENT);
+	uh->skip_mode.skip_mode_frame[0] =
+		ctrl_fh->skip_mode_frame[0] - V4L2_AV1_REF_LAST_FRAME;
+	uh->skip_mode.skip_mode_frame[1] =
+		ctrl_fh->skip_mode_frame[1] - V4L2_AV1_REF_LAST_FRAME;
+	uh->skip_mode.skip_mode_allowed = ctrl_fh->skip_mode_frame[0] ? 1 : 0;
+
+	vdec_av1_slice_setup_seg(&uh->seg, &ctrl_fh->segmentation);
+	uh->delta_q_lf.delta_q_present = QUANT_FLAG(&ctrl_fh->quantization, DELTA_Q_PRESENT);
+	uh->delta_q_lf.delta_q_res = 1 << ctrl_fh->quantization.delta_q_res;
+	uh->delta_q_lf.delta_lf_present =
+		BIT_FLAG(&ctrl_fh->loop_filter, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_LF_PRESENT);
+	uh->delta_q_lf.delta_lf_res = ctrl_fh->loop_filter.delta_lf_res;
+	uh->delta_q_lf.delta_lf_multi =
+		BIT_FLAG(&ctrl_fh->loop_filter, V4L2_AV1_LOOP_FILTER_FLAG_DELTA_LF_MULTI);
+	vdec_av1_slice_setup_quant(&uh->quant, &ctrl_fh->quantization);
+
+	uh->coded_loss_less = 1;
+	for (i = 0; i < V4L2_AV1_MAX_SEGMENTS; i++) {
+		uh->quant.qindex[i] = vdec_av1_slice_get_qindex(uh, i);
+		uh->loss_less_array[i] =
+			(uh->quant.qindex[i] == 0 && uh->quant.delta_qydc == 0 &&
+			uh->quant.delta_quac == 0 && uh->quant.delta_qudc == 0 &&
+			uh->quant.delta_qvac == 0 && uh->quant.delta_qvdc == 0);
+
+		if (!uh->loss_less_array[i])
+			uh->coded_loss_less = 0;
+	}
+
+	vdec_av1_slice_setup_lr(&uh->lr, &ctrl_fh->loop_restoration);
+	uh->superres_denom = ctrl_fh->superres_denom;
+	vdec_av1_slice_setup_lf(&uh->loop_filter, &ctrl_fh->loop_filter);
+	vdec_av1_slice_setup_cdef(&uh->cdef, &ctrl_fh->cdef);
+	vdec_av1_slice_setup_tile(frame, &ctrl_fh->tile_info);
+}
+
+static int vdec_av1_slice_setup_tile_group(struct vdec_av1_slice_instance *instance,
+					   struct vdec_av1_slice_vsi *vsi)
+{
+	struct v4l2_ctrl_av1_tile_group_entry *ctrl_tge;
+	struct vdec_av1_slice_tile_group *tile_group = &instance->tile_group;
+	struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh;
+	struct vdec_av1_slice_tile *tile = &uh->tile;
+	struct v4l2_ctrl *ctrl;
+	u32 tge_size;
+	int i;
+
+	ctrl = v4l2_ctrl_find(&instance->ctx->ctrl_hdl, V4L2_CID_STATELESS_AV1_TILE_GROUP_ENTRY);
+	if (!ctrl)
+		return -EINVAL;
+
+	tge_size = ctrl->elems;
+	ctrl_tge = (struct v4l2_ctrl_av1_tile_group_entry *)ctrl->p_cur.p;
+
+	tile_group->num_tiles = tile->tile_cols * tile->tile_rows;
+
+	if (tile_group->num_tiles != tge_size ||
+	    tile_group->num_tiles > V4L2_AV1_MAX_TILE_COUNT) {
+		mtk_vdec_err(instance->ctx, "invalid tge_size %d, tile_num:%d\n",
+			     tge_size, tile_group->num_tiles);
+		return -EINVAL;
+	}
+
+	for (i = 0; i < tge_size; i++) {
+		if (i != ctrl_tge[i].tile_row * vsi->frame.uh.tile.tile_cols +
+		    ctrl_tge[i].tile_col) {
+			mtk_vdec_err(instance->ctx, "invalid tge info %d, %d %d %d\n",
+				     i, ctrl_tge[i].tile_row, ctrl_tge[i].tile_col,
+				     vsi->frame.uh.tile.tile_rows);
+			return -EINVAL;
+		}
+		tile_group->tile_size[i] = ctrl_tge[i].tile_size;
+		tile_group->tile_start_offset[i] = ctrl_tge[i].tile_offset;
+	}
+
+	return 0;
+}
+
+static inline void vdec_av1_slice_setup_state(struct vdec_av1_slice_vsi *vsi)
+{
+	memset(&vsi->state, 0, sizeof(vsi->state));
+}
+
+static void vdec_av1_slice_setup_scale_factors(struct vdec_av1_slice_frame_refs *frame_ref,
+					       struct vdec_av1_slice_frame_info *ref_frame_info,
+					       struct vdec_av1_slice_uncompressed_header *uh)
+{
+	struct vdec_av1_slice_scale_factors *scale_factors = &frame_ref->scale_factors;
+	u32 ref_upscaled_width = ref_frame_info->upscaled_width;
+	u32 ref_frame_height = ref_frame_info->frame_height;
+	u32 frame_width = uh->frame_width;
+	u32 frame_height = uh->frame_height;
+
+	if (!vdec_av1_slice_need_scale(ref_upscaled_width, ref_frame_height,
+				       frame_width, frame_height)) {
+		scale_factors->x_scale = -1;
+		scale_factors->y_scale = -1;
+		scale_factors->is_scaled = 0;
+		return;
+	}
+
+	scale_factors->x_scale =
+		((ref_upscaled_width << AV1_REF_SCALE_SHIFT) + (frame_width >> 1)) / frame_width;
+	scale_factors->y_scale =
+		((ref_frame_height << AV1_REF_SCALE_SHIFT) + (frame_height >> 1)) / frame_height;
+	scale_factors->is_scaled =
+		(scale_factors->x_scale != AV1_REF_INVALID_SCALE) &&
+		(scale_factors->y_scale != AV1_REF_INVALID_SCALE) &&
+		(scale_factors->x_scale != AV1_REF_NO_SCALE ||
+		 scale_factors->y_scale != AV1_REF_NO_SCALE);
+	scale_factors->x_step =
+		AV1_DIV_ROUND_UP_POW2(scale_factors->x_scale,
+				      AV1_REF_SCALE_SHIFT - AV1_SCALE_SUBPEL_BITS);
+	scale_factors->y_step =
+		AV1_DIV_ROUND_UP_POW2(scale_factors->y_scale,
+				      AV1_REF_SCALE_SHIFT - AV1_SCALE_SUBPEL_BITS);
+}
+
+static unsigned char vdec_av1_slice_get_sign_bias(int a,
+						  int b,
+						  u8 enable_order_hint,
+						  u8 order_hint_bits)
+{
+	int diff = 0;
+	int m = 0;
+	unsigned char result = 0;
+
+	if (!enable_order_hint)
+		return 0;
+
+	diff = a - b;
+	m = 1 << (order_hint_bits - 1);
+	diff = (diff & (m - 1)) - (diff & m);
+
+	if (diff > 0)
+		result = 1;
+
+	return result;
+}
+
+static void vdec_av1_slice_setup_ref(struct vdec_av1_slice_pfc *pfc,
+				     struct v4l2_ctrl_av1_frame *ctrl_fh)
+{
+	struct vdec_av1_slice_vsi *vsi = &pfc->vsi;
+	struct vdec_av1_slice_frame *frame = &vsi->frame;
+	struct vdec_av1_slice_slot *slots = &vsi->slots;
+	struct vdec_av1_slice_uncompressed_header *uh = &frame->uh;
+	struct vdec_av1_slice_seq_header *seq = &frame->seq;
+	struct vdec_av1_slice_frame_info *cur_frame_info =
+		&slots->frame_info[vsi->slot_id];
+	struct vdec_av1_slice_frame_info *frame_info;
+	int i, slot_id;
+
+	if (uh->frame_is_intra)
+		return;
+
+	for (i = 0; i < V4L2_AV1_REFS_PER_FRAME; i++) {
+		int ref_idx = ctrl_fh->ref_frame_idx[i];
+
+		pfc->ref_idx[i] = ctrl_fh->reference_frame_ts[ref_idx];
+		slot_id = frame->ref_frame_map[ref_idx];
+		frame_info = &slots->frame_info[slot_id];
+		if (slot_id == AV1_INVALID_IDX) {
+			pr_err(MTK_DBG_V4L2_STR "cannot match reference[%d] 0x%llx\n", i,
+			       ctrl_fh->reference_frame_ts[ref_idx]);
+			frame->order_hints[i] = 0;
+			frame->ref_frame_valid[i] = 0;
+			continue;
+		}
+
+		frame->frame_refs[i].ref_fb_idx = slot_id;
+		vdec_av1_slice_setup_scale_factors(&frame->frame_refs[i],
+						   frame_info, uh);
+		if (!seq->enable_order_hint)
+			frame->ref_frame_sign_bias[i + 1] = 0;
+		else
+			frame->ref_frame_sign_bias[i + 1] =
+				vdec_av1_slice_get_sign_bias(frame_info->order_hint,
+							     uh->order_hint,
+							     seq->enable_order_hint,
+							     seq->order_hint_bits);
+
+		frame->order_hints[i] = ctrl_fh->order_hints[i + 1];
+		cur_frame_info->order_hints[i] = frame->order_hints[i];
+		frame->ref_frame_valid[i] = 1;
+	}
+}
+
+static void vdec_av1_slice_get_previous(struct vdec_av1_slice_vsi *vsi)
+{
+	struct vdec_av1_slice_frame *frame = &vsi->frame;
+
+	if (frame->uh.primary_ref_frame == AV1_PRIMARY_REF_NONE)
+		frame->prev_fb_idx = AV1_INVALID_IDX;
+	else
+		frame->prev_fb_idx = frame->frame_refs[frame->uh.primary_ref_frame].ref_fb_idx;
+}
+
+static inline void vdec_av1_slice_setup_operating_mode(struct vdec_av1_slice_instance *instance,
+						       struct vdec_av1_slice_frame *frame)
+{
+	frame->large_scale_tile = 0;
+}
+
+static int vdec_av1_slice_setup_pfc(struct vdec_av1_slice_instance *instance,
+				    struct vdec_av1_slice_pfc *pfc)
+{
+	struct v4l2_ctrl_av1_frame *ctrl_fh;
+	struct v4l2_ctrl_av1_sequence *ctrl_seq;
+	struct vdec_av1_slice_vsi *vsi = &pfc->vsi;
+	int ret = 0;
+
+	/* frame header */
+	ctrl_fh = (struct v4l2_ctrl_av1_frame *)
+		  vdec_av1_get_ctrl_ptr(instance->ctx,
+					V4L2_CID_STATELESS_AV1_FRAME);
+	if (IS_ERR(ctrl_fh))
+		return PTR_ERR(ctrl_fh);
+
+	ctrl_seq = (struct v4l2_ctrl_av1_sequence *)
+		   vdec_av1_get_ctrl_ptr(instance->ctx,
+					 V4L2_CID_STATELESS_AV1_SEQUENCE);
+	if (IS_ERR(ctrl_seq))
+		return PTR_ERR(ctrl_seq);
+
+	/* setup vsi information */
+	vdec_av1_slice_setup_seq(&vsi->frame.seq, ctrl_seq);
+	vdec_av1_slice_setup_uh(instance, &vsi->frame, ctrl_fh);
+	vdec_av1_slice_setup_operating_mode(instance, &vsi->frame);
+
+	vdec_av1_slice_setup_state(vsi);
+	vdec_av1_slice_setup_slot(instance, vsi, ctrl_fh);
+	vdec_av1_slice_setup_ref(pfc, ctrl_fh);
+	vdec_av1_slice_get_previous(vsi);
+
+	pfc->seq = instance->seq;
+	instance->seq++;
+
+	return ret;
+}
+
+static void vdec_av1_slice_setup_lat_buffer(struct vdec_av1_slice_instance *instance,
+					    struct vdec_av1_slice_vsi *vsi,
+					    struct mtk_vcodec_mem *bs,
+					    struct vdec_lat_buf *lat_buf)
+{
+	struct vdec_av1_slice_work_buffer *work_buffer;
+	int i;
+
+	vsi->bs.dma_addr = bs->dma_addr;
+	vsi->bs.size = bs->size;
+
+	vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size;
+	vsi->trans.dma_addr = lat_buf->ctx->msg_queue.wdma_wptr_addr;
+	/* used to store trans end */
+	vsi->trans.dma_addr_end = lat_buf->ctx->msg_queue.wdma_rptr_addr;
+	vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr;
+	vsi->err_map.size = lat_buf->wdma_err_addr.size;
+	vsi->rd_mv.dma_addr = lat_buf->rd_mv_addr.dma_addr;
+	vsi->rd_mv.size = lat_buf->rd_mv_addr.size;
+
+	vsi->row_info.buf = 0;
+	vsi->row_info.size = 0;
+
+	work_buffer = vsi->work_buffer;
+
+	for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++) {
+		work_buffer[i].mv_addr.buf = instance->mv[i].dma_addr;
+		work_buffer[i].mv_addr.size = instance->mv[i].size;
+		work_buffer[i].segid_addr.buf = instance->seg[i].dma_addr;
+		work_buffer[i].segid_addr.size = instance->seg[i].size;
+		work_buffer[i].cdf_addr.buf = instance->cdf[i].dma_addr;
+		work_buffer[i].cdf_addr.size = instance->cdf[i].size;
+	}
+
+	vsi->cdf_tmp.buf = instance->cdf_temp.dma_addr;
+	vsi->cdf_tmp.size = instance->cdf_temp.size;
+
+	vsi->tile.buf = instance->tile.dma_addr;
+	vsi->tile.size = instance->tile.size;
+	memcpy(lat_buf->tile_addr.va, instance->tile.va, 64 * instance->tile_group.num_tiles);
+
+	vsi->cdf_table.buf = instance->cdf_table.dma_addr;
+	vsi->cdf_table.size = instance->cdf_table.size;
+	vsi->iq_table.buf = instance->iq_table.dma_addr;
+	vsi->iq_table.size = instance->iq_table.size;
+}
+
+static void vdec_av1_slice_setup_seg_buffer(struct vdec_av1_slice_instance *instance,
+					    struct vdec_av1_slice_vsi *vsi)
+{
+	struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh;
+	struct mtk_vcodec_mem *buf;
+
+	/* reset segment buffer */
+	if (uh->primary_ref_frame == AV1_PRIMARY_REF_NONE || !uh->seg.segmentation_enabled) {
+		mtk_vdec_debug(instance->ctx, "reset seg %d\n", vsi->slot_id);
+		if (vsi->slot_id != AV1_INVALID_IDX) {
+			buf = &instance->seg[vsi->slot_id];
+			memset(buf->va, 0, buf->size);
+		}
+	}
+}
+
+static void vdec_av1_slice_setup_tile_buffer(struct vdec_av1_slice_instance *instance,
+					     struct vdec_av1_slice_vsi *vsi,
+					     struct mtk_vcodec_mem *bs)
+{
+	struct vdec_av1_slice_tile_group *tile_group = &instance->tile_group;
+	struct vdec_av1_slice_uncompressed_header *uh = &vsi->frame.uh;
+	struct vdec_av1_slice_tile *tile = &uh->tile;
+	u32 tile_num, tile_row, tile_col;
+	u32 allow_update_cdf = 0;
+	u32 sb_boundary_x_m1 = 0, sb_boundary_y_m1 = 0;
+	int tile_info_base;
+	u64 tile_buf_pa;
+	u32 *tile_info_buf = instance->tile.va;
+	u64 pa = (u64)bs->dma_addr;
+
+	if (uh->disable_cdf_update == 0)
+		allow_update_cdf = 1;
+
+	for (tile_num = 0; tile_num < tile_group->num_tiles; tile_num++) {
+		/* each uint32 takes place of 4 bytes */
+		tile_info_base = (AV1_TILE_BUF_SIZE * tile_num) >> 2;
+		tile_row = tile_num / tile->tile_cols;
+		tile_col = tile_num % tile->tile_cols;
+		tile_info_buf[tile_info_base + 0] = (tile_group->tile_size[tile_num] << 3);
+		tile_buf_pa = pa + tile_group->tile_start_offset[tile_num];
+
+		/* save av1 tile high 4bits(bit 32-35) address in lower 4 bits position
+		 * and clear original for hw requirement.
+		 */
+		tile_info_buf[tile_info_base + 1] = (tile_buf_pa & 0xFFFFFFF0ull) |
+			((tile_buf_pa & 0xF00000000ull) >> 32);
+		tile_info_buf[tile_info_base + 2] = (tile_buf_pa & 0xFull) << 3;
+
+		sb_boundary_x_m1 =
+			(tile->mi_col_starts[tile_col + 1] - tile->mi_col_starts[tile_col] - 1) &
+			0x3f;
+		sb_boundary_y_m1 =
+			(tile->mi_row_starts[tile_row + 1] - tile->mi_row_starts[tile_row] - 1) &
+			0x1ff;
+
+		tile_info_buf[tile_info_base + 3] = (sb_boundary_y_m1 << 7) | sb_boundary_x_m1;
+		tile_info_buf[tile_info_base + 4] = ((allow_update_cdf << 18) | (1 << 16));
+
+		if (tile_num == tile->context_update_tile_id &&
+		    uh->disable_frame_end_update_cdf == 0)
+			tile_info_buf[tile_info_base + 4] |= (1 << 17);
+
+		mtk_vdec_debug(instance->ctx, "// tile buf %d pos(%dx%d) offset 0x%x\n",
+			       tile_num, tile_row, tile_col, tile_info_base);
+		mtk_vdec_debug(instance->ctx, "// %08x %08x %08x %08x\n",
+			       tile_info_buf[tile_info_base + 0],
+			       tile_info_buf[tile_info_base + 1],
+			       tile_info_buf[tile_info_base + 2],
+			       tile_info_buf[tile_info_base + 3]);
+		mtk_vdec_debug(instance->ctx, "// %08x %08x %08x %08x\n",
+			       tile_info_buf[tile_info_base + 4],
+			       tile_info_buf[tile_info_base + 5],
+			       tile_info_buf[tile_info_base + 6],
+			       tile_info_buf[tile_info_base + 7]);
+	}
+}
+
+static int vdec_av1_slice_setup_lat(struct vdec_av1_slice_instance *instance,
+				    struct mtk_vcodec_mem *bs,
+				    struct vdec_lat_buf *lat_buf,
+				    struct vdec_av1_slice_pfc *pfc)
+{
+	struct vdec_av1_slice_vsi *vsi = &pfc->vsi;
+	int ret;
+
+	ret = vdec_av1_slice_setup_lat_from_src_buf(instance, vsi, lat_buf);
+	if (ret)
+		return ret;
+
+	ret = vdec_av1_slice_setup_pfc(instance, pfc);
+	if (ret)
+		return ret;
+
+	ret = vdec_av1_slice_setup_tile_group(instance, vsi);
+	if (ret)
+		return ret;
+
+	ret = vdec_av1_slice_alloc_working_buffer(instance, vsi);
+	if (ret)
+		return ret;
+
+	vdec_av1_slice_setup_seg_buffer(instance, vsi);
+	vdec_av1_slice_setup_tile_buffer(instance, vsi, bs);
+	vdec_av1_slice_setup_lat_buffer(instance, vsi, bs, lat_buf);
+
+	return 0;
+}
+
+static int vdec_av1_slice_update_lat(struct vdec_av1_slice_instance *instance,
+				     struct vdec_lat_buf *lat_buf,
+				     struct vdec_av1_slice_pfc *pfc)
+{
+	struct vdec_av1_slice_vsi *vsi;
+
+	vsi = &pfc->vsi;
+	mtk_vdec_debug(instance->ctx, "frame %u LAT CRC 0x%08x, output size is %d\n",
+		       pfc->seq, vsi->state.crc[0], vsi->state.out_size);
+
+	/* buffer full, need to re-decode */
+	if (vsi->state.full) {
+		/* buffer not enough */
+		if (vsi->trans.dma_addr_end - vsi->trans.dma_addr == vsi->ube.size)
+			return -ENOMEM;
+		return -EAGAIN;
+	}
+
+	instance->width = vsi->frame.uh.upscaled_width;
+	instance->height = vsi->frame.uh.frame_height;
+	instance->frame_type = vsi->frame.uh.frame_type;
+
+	return 0;
+}
+
+static int vdec_av1_slice_setup_core_to_dst_buf(struct vdec_av1_slice_instance *instance,
+						struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_v4l2_buffer *dst;
+
+	dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx);
+	if (!dst)
+		return -EINVAL;
+
+	v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, dst, true);
+
+	return 0;
+}
+
+static int vdec_av1_slice_setup_core_buffer(struct vdec_av1_slice_instance *instance,
+					    struct vdec_av1_slice_pfc *pfc,
+					    struct vdec_av1_slice_vsi *vsi,
+					    struct vdec_fb *fb,
+					    struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_buffer *vb;
+	struct vb2_queue *vq;
+	int w, h, plane, size;
+	int i;
+
+	plane = instance->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes;
+	w = vsi->frame.uh.upscaled_width;
+	h = vsi->frame.uh.frame_height;
+	size = ALIGN(w, VCODEC_DEC_ALIGNED_64) * ALIGN(h, VCODEC_DEC_ALIGNED_64);
+
+	/* frame buffer */
+	vsi->fb.y.dma_addr = fb->base_y.dma_addr;
+	if (plane == 1)
+		vsi->fb.c.dma_addr = fb->base_y.dma_addr + size;
+	else
+		vsi->fb.c.dma_addr = fb->base_c.dma_addr;
+
+	/* reference buffers */
+	vq = v4l2_m2m_get_vq(instance->ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	if (!vq)
+		return -EINVAL;
+
+	/* get current output buffer */
+	vb = &v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx)->vb2_buf;
+	if (!vb)
+		return -EINVAL;
+
+	/* get buffer address from vb2buf */
+	for (i = 0; i < V4L2_AV1_REFS_PER_FRAME; i++) {
+		struct vdec_av1_slice_fb *vref = &vsi->ref[i];
+
+		vb = vb2_find_buffer(vq, pfc->ref_idx[i]);
+		if (!vb) {
+			memset(vref, 0, sizeof(*vref));
+			continue;
+		}
+
+		vref->y.dma_addr = vb2_dma_contig_plane_dma_addr(vb, 0);
+		if (plane == 1)
+			vref->c.dma_addr = vref->y.dma_addr + size;
+		else
+			vref->c.dma_addr = vb2_dma_contig_plane_dma_addr(vb, 1);
+	}
+	vsi->tile.dma_addr = lat_buf->tile_addr.dma_addr;
+	vsi->tile.size = lat_buf->tile_addr.size;
+
+	return 0;
+}
+
+static int vdec_av1_slice_setup_core(struct vdec_av1_slice_instance *instance,
+				     struct vdec_fb *fb,
+				     struct vdec_lat_buf *lat_buf,
+				     struct vdec_av1_slice_pfc *pfc)
+{
+	struct vdec_av1_slice_vsi *vsi = &pfc->vsi;
+	int ret;
+
+	ret = vdec_av1_slice_setup_core_to_dst_buf(instance, lat_buf);
+	if (ret)
+		return ret;
+
+	ret = vdec_av1_slice_setup_core_buffer(instance, pfc, vsi, fb, lat_buf);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int vdec_av1_slice_update_core(struct vdec_av1_slice_instance *instance,
+				      struct vdec_lat_buf *lat_buf,
+				      struct vdec_av1_slice_pfc *pfc)
+{
+	struct vdec_av1_slice_vsi *vsi = instance->core_vsi;
+
+	mtk_vdec_debug(instance->ctx, "frame %u Y_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[0], vsi->state.crc[1],
+		       vsi->state.crc[2], vsi->state.crc[3]);
+	mtk_vdec_debug(instance->ctx, "frame %u C_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[8], vsi->state.crc[9],
+		       vsi->state.crc[10], vsi->state.crc[11]);
+
+	return 0;
+}
+
+static int vdec_av1_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_av1_slice_instance *instance;
+	struct vdec_av1_slice_init_vsi *vsi;
+	int ret;
+
+	instance = kzalloc(sizeof(*instance), GFP_KERNEL);
+	if (!instance)
+		return -ENOMEM;
+
+	instance->ctx = ctx;
+	instance->vpu.id = SCP_IPI_VDEC_LAT;
+	instance->vpu.core_id = SCP_IPI_VDEC_CORE;
+	instance->vpu.ctx = ctx;
+	instance->vpu.codec_type = ctx->current_codec;
+
+	ret = vpu_dec_init(&instance->vpu);
+	if (ret) {
+		mtk_vdec_err(ctx, "failed to init vpu dec, ret %d\n", ret);
+		goto error_vpu_init;
+	}
+
+	/* init vsi and global flags */
+	vsi = instance->vpu.vsi;
+	if (!vsi) {
+		mtk_vdec_err(ctx, "failed to get AV1 vsi\n");
+		ret = -EINVAL;
+		goto error_vsi;
+	}
+	instance->init_vsi = vsi;
+	instance->core_vsi = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler, (u32)vsi->core_vsi);
+
+	if (!instance->core_vsi) {
+		mtk_vdec_err(ctx, "failed to get AV1 core vsi\n");
+		ret = -EINVAL;
+		goto error_vsi;
+	}
+
+	if (vsi->vsi_size != sizeof(struct vdec_av1_slice_vsi))
+		mtk_vdec_err(ctx, "remote vsi size 0x%x mismatch! expected: 0x%zx\n",
+			     vsi->vsi_size, sizeof(struct vdec_av1_slice_vsi));
+
+	instance->irq_enabled = 1;
+	instance->inneracing_mode = IS_VDEC_INNER_RACING(instance->ctx->dev->dec_capability);
+
+	mtk_vdec_debug(ctx, "vsi 0x%p core_vsi 0x%llx 0x%p, inneracing_mode %d\n",
+		       vsi, vsi->core_vsi, instance->core_vsi, instance->inneracing_mode);
+
+	ret = vdec_av1_slice_init_cdf_table(instance);
+	if (ret)
+		goto error_vsi;
+
+	ret = vdec_av1_slice_init_iq_table(instance);
+	if (ret)
+		goto error_vsi;
+
+	ctx->drv_handle = instance;
+
+	return 0;
+error_vsi:
+	vpu_dec_deinit(&instance->vpu);
+error_vpu_init:
+	kfree(instance);
+
+	return ret;
+}
+
+static void vdec_av1_slice_deinit(void *h_vdec)
+{
+	struct vdec_av1_slice_instance *instance = h_vdec;
+
+	if (!instance)
+		return;
+	mtk_vdec_debug(instance->ctx, "h_vdec 0x%p\n", h_vdec);
+	vpu_dec_deinit(&instance->vpu);
+	vdec_av1_slice_free_working_buffer(instance);
+	vdec_msg_queue_deinit(&instance->ctx->msg_queue, instance->ctx);
+	kfree(instance);
+}
+
+static int vdec_av1_slice_flush(void *h_vdec, struct mtk_vcodec_mem *bs,
+				struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_av1_slice_instance *instance = h_vdec;
+	int i;
+
+	mtk_vdec_debug(instance->ctx, "flush ...\n");
+
+	vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue);
+
+	for (i = 0; i < AV1_MAX_FRAME_BUF_COUNT; i++)
+		vdec_av1_slice_clear_fb(&instance->slots.frame_info[i]);
+
+	return vpu_dec_reset(&instance->vpu);
+}
+
+static void vdec_av1_slice_get_pic_info(struct vdec_av1_slice_instance *instance)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	u32 data[3];
+
+	mtk_vdec_debug(ctx, "w %u h %u\n", ctx->picinfo.pic_w, ctx->picinfo.pic_h);
+
+	data[0] = ctx->picinfo.pic_w;
+	data[1] = ctx->picinfo.pic_h;
+	data[2] = ctx->capture_fourcc;
+	vpu_dec_get_param(&instance->vpu, data, 3, GET_PARAM_PIC_INFO);
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.fb_sz[0] = instance->vpu.fb_sz[0];
+	ctx->picinfo.fb_sz[1] = instance->vpu.fb_sz[1];
+}
+
+static inline void vdec_av1_slice_get_dpb_size(struct vdec_av1_slice_instance *instance,
+					       u32 *dpb_sz)
+{
+	/* refer av1 specification */
+	*dpb_sz = V4L2_AV1_TOTAL_REFS_PER_FRAME + 1;
+}
+
+static void vdec_av1_slice_get_crop_info(struct vdec_av1_slice_instance *instance,
+					 struct v4l2_rect *cr)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+
+	cr->left = 0;
+	cr->top = 0;
+	cr->width = ctx->picinfo.pic_w;
+	cr->height = ctx->picinfo.pic_h;
+
+	mtk_vdec_debug(ctx, "l=%d, t=%d, w=%d, h=%d\n",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static int vdec_av1_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out)
+{
+	struct vdec_av1_slice_instance *instance = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		vdec_av1_slice_get_pic_info(instance);
+		break;
+	case GET_PARAM_DPB_SIZE:
+		vdec_av1_slice_get_dpb_size(instance, out);
+		break;
+	case GET_PARAM_CROP_INFO:
+		vdec_av1_slice_get_crop_info(instance, out);
+		break;
+	default:
+		mtk_vdec_err(instance->ctx, "invalid get parameter type=%d\n", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int vdec_av1_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				     struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_av1_slice_instance *instance = h_vdec;
+	struct vdec_lat_buf *lat_buf;
+	struct vdec_av1_slice_pfc *pfc;
+	struct vdec_av1_slice_vsi *vsi;
+	struct mtk_vcodec_dec_ctx *ctx;
+	int ret;
+
+	if (!instance || !instance->ctx)
+		return -EINVAL;
+
+	ctx = instance->ctx;
+	/* init msgQ for the first time */
+	if (vdec_msg_queue_init(&ctx->msg_queue, ctx,
+				vdec_av1_slice_core_decode, sizeof(*pfc))) {
+		mtk_vdec_err(ctx, "failed to init AV1 msg queue\n");
+		return -ENOMEM;
+	}
+
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vdec_av1_slice_flush(h_vdec, bs, fb, res_chg);
+
+	lat_buf = vdec_msg_queue_dqbuf(&ctx->msg_queue.lat_ctx);
+	if (!lat_buf) {
+		mtk_vdec_err(ctx, "failed to get AV1 lat buf\n");
+		return -EAGAIN;
+	}
+	pfc = (struct vdec_av1_slice_pfc *)lat_buf->private_data;
+	if (!pfc) {
+		ret = -EINVAL;
+		goto err_free_fb_out;
+	}
+	vsi = &pfc->vsi;
+
+	ret = vdec_av1_slice_setup_lat(instance, bs, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "failed to setup AV1 lat ret %d\n", ret);
+		goto err_free_fb_out;
+	}
+
+	vdec_av1_slice_vsi_to_remote(vsi, instance->vsi);
+	ret = vpu_dec_start(&instance->vpu, NULL, 0);
+	if (ret) {
+		mtk_vdec_err(ctx, "failed to dec AV1 ret %d\n", ret);
+		goto err_free_fb_out;
+	}
+	if (instance->inneracing_mode)
+		vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf);
+
+	if (instance->irq_enabled) {
+		ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED,
+						   WAIT_INTR_TIMEOUT_MS,
+						   MTK_VDEC_LAT0);
+		/* update remote vsi if decode timeout */
+		if (ret) {
+			mtk_vdec_err(ctx, "AV1 Frame %d decode timeout %d\n", pfc->seq, ret);
+			WRITE_ONCE(instance->vsi->state.timeout, 1);
+		}
+		vpu_dec_end(&instance->vpu);
+	}
+
+	vdec_av1_slice_vsi_from_remote(vsi, instance->vsi);
+	ret = vdec_av1_slice_update_lat(instance, lat_buf, pfc);
+
+	/* LAT trans full, re-decode */
+	if (ret == -EAGAIN) {
+		mtk_vdec_err(ctx, "AV1 Frame %d trans full\n", pfc->seq);
+		if (!instance->inneracing_mode)
+			vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf);
+		return 0;
+	}
+
+	/* LAT trans full, no more UBE or decode timeout */
+	if (ret == -ENOMEM || vsi->state.timeout) {
+		mtk_vdec_err(ctx, "AV1 Frame %d insufficient buffer or timeout\n", pfc->seq);
+		if (!instance->inneracing_mode)
+			vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf);
+		return -EBUSY;
+	}
+	vsi->trans.dma_addr_end += ctx->msg_queue.wdma_addr.dma_addr;
+	mtk_vdec_debug(ctx, "lat dma 1 0x%pad 0x%pad\n",
+		       &pfc->vsi.trans.dma_addr, &pfc->vsi.trans.dma_addr_end);
+
+	vdec_msg_queue_update_ube_wptr(&ctx->msg_queue, vsi->trans.dma_addr_end);
+
+	if (!instance->inneracing_mode)
+		vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf);
+	memcpy(&instance->slots, &vsi->slots, sizeof(instance->slots));
+
+	return 0;
+
+err_free_fb_out:
+	vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf);
+
+	if (pfc)
+		mtk_vdec_err(ctx, "slice dec number: %d err: %d", pfc->seq, ret);
+
+	return ret;
+}
+
+static int vdec_av1_slice_core_decode(struct vdec_lat_buf *lat_buf)
+{
+	struct vdec_av1_slice_instance *instance;
+	struct vdec_av1_slice_pfc *pfc;
+	struct mtk_vcodec_dec_ctx *ctx = NULL;
+	struct vdec_fb *fb = NULL;
+	int ret = -EINVAL;
+
+	if (!lat_buf)
+		return -EINVAL;
+
+	pfc = lat_buf->private_data;
+	ctx = lat_buf->ctx;
+	if (!pfc || !ctx)
+		return -EINVAL;
+
+	instance = ctx->drv_handle;
+	if (!instance)
+		goto err;
+
+	fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx);
+	if (!fb) {
+		ret = -EBUSY;
+		goto err;
+	}
+
+	ret = vdec_av1_slice_setup_core(instance, fb, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "vdec_av1_slice_setup_core\n");
+		goto err;
+	}
+	vdec_av1_slice_vsi_to_remote(&pfc->vsi, instance->core_vsi);
+	ret = vpu_dec_core(&instance->vpu);
+	if (ret) {
+		mtk_vdec_err(ctx, "vpu_dec_core\n");
+		goto err;
+	}
+
+	if (instance->irq_enabled) {
+		ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED,
+						   WAIT_INTR_TIMEOUT_MS,
+						   MTK_VDEC_CORE);
+		/* update remote vsi if decode timeout */
+		if (ret) {
+			mtk_vdec_err(ctx, "AV1 frame %d core timeout\n", pfc->seq);
+			WRITE_ONCE(instance->vsi->state.timeout, 1);
+		}
+		vpu_dec_core_end(&instance->vpu);
+	}
+
+	ret = vdec_av1_slice_update_core(instance, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "vdec_av1_slice_update_core\n");
+		goto err;
+	}
+
+	mtk_vdec_debug(ctx, "core dma_addr_end 0x%pad\n",
+		       &instance->core_vsi->trans.dma_addr_end);
+	vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, instance->core_vsi->trans.dma_addr_end);
+
+	ctx->dev->vdec_pdata->cap_to_disp(ctx, 0, lat_buf->src_buf_req);
+
+	return 0;
+
+err:
+	/* always update read pointer */
+	vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end);
+
+	if (fb)
+		ctx->dev->vdec_pdata->cap_to_disp(ctx, 1, lat_buf->src_buf_req);
+
+	return ret;
+}
+
+const struct vdec_common_if vdec_av1_slice_lat_if = {
+	.init		= vdec_av1_slice_init,
+	.decode		= vdec_av1_slice_lat_decode,
+	.get_param	= vdec_av1_slice_get_param,
+	.deinit		= vdec_av1_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c
new file mode 100644
index 0000000000..bf7dffe60d
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_if.c
@@ -0,0 +1,494 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+
+#include "../vdec_drv_if.h"
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_vpu_if.h"
+#include "../vdec_drv_base.h"
+
+#define NAL_NON_IDR_SLICE			0x01
+#define NAL_IDR_SLICE				0x05
+#define NAL_H264_PPS				0x08
+#define NAL_TYPE(value)				((value) & 0x1F)
+
+#define BUF_PREDICTION_SZ			(32 * 1024)
+
+#define MB_UNIT_LEN				16
+
+/* motion vector size (bytes) for every macro block */
+#define HW_MB_STORE_SZ				64
+
+#define H264_MAX_FB_NUM				17
+#define HDR_PARSING_BUF_SZ			1024
+
+#define DEC_ERR_RET(ret)			((ret) >> 16)
+#define H264_ERR_NOT_VALID			3
+
+/**
+ * struct h264_fb - h264 decode frame buffer information
+ * @vdec_fb_va  : virtual address of struct vdec_fb
+ * @y_fb_dma    : dma address of Y frame buffer (luma)
+ * @c_fb_dma    : dma address of C frame buffer (chroma)
+ * @poc         : picture order count of frame buffer
+ * @reserved    : for 8 bytes alignment
+ */
+struct h264_fb {
+	uint64_t vdec_fb_va;
+	uint64_t y_fb_dma;
+	uint64_t c_fb_dma;
+	int32_t poc;
+	uint32_t reserved;
+};
+
+/**
+ * struct h264_ring_fb_list - ring frame buffer list
+ * @fb_list   : frame buffer array
+ * @read_idx  : read index
+ * @write_idx : write index
+ * @count     : buffer count in list
+ * @reserved  : for 8 bytes alignment
+ */
+struct h264_ring_fb_list {
+	struct h264_fb fb_list[H264_MAX_FB_NUM];
+	unsigned int read_idx;
+	unsigned int write_idx;
+	unsigned int count;
+	unsigned int reserved;
+};
+
+/**
+ * struct vdec_h264_dec_info - decode information
+ * @dpb_sz		: decoding picture buffer size
+ * @resolution_changed  : resolution change happen
+ * @realloc_mv_buf	: flag to notify driver to re-allocate mv buffer
+ * @reserved		: for 8 bytes alignment
+ * @bs_dma		: Input bit-stream buffer dma address
+ * @y_fb_dma		: Y frame buffer dma address
+ * @c_fb_dma		: C frame buffer dma address
+ * @vdec_fb_va		: VDEC frame buffer struct virtual address
+ */
+struct vdec_h264_dec_info {
+	uint32_t dpb_sz;
+	uint32_t resolution_changed;
+	uint32_t realloc_mv_buf;
+	uint32_t reserved;
+	uint64_t bs_dma;
+	uint64_t y_fb_dma;
+	uint64_t c_fb_dma;
+	uint64_t vdec_fb_va;
+};
+
+/**
+ * struct vdec_h264_vsi - shared memory for decode information exchange
+ *                        between VPU and Host.
+ *                        The memory is allocated by VPU then mapping to Host
+ *                        in vpu_dec_init() and freed in vpu_dec_deinit()
+ *                        by VPU.
+ *                        AP-W/R : AP is writer/reader on this item
+ *                        VPU-W/R: VPU is write/reader on this item
+ * @hdr_buf      : Header parsing buffer (AP-W, VPU-R)
+ * @pred_buf_dma : HW working predication buffer dma address (AP-W, VPU-R)
+ * @mv_buf_dma   : HW working motion vector buffer dma address (AP-W, VPU-R)
+ * @list_free    : free frame buffer ring list (AP-W/R, VPU-W)
+ * @list_disp    : display frame buffer ring list (AP-R, VPU-W)
+ * @dec          : decode information (AP-R, VPU-W)
+ * @pic          : picture information (AP-R, VPU-W)
+ * @crop         : crop information (AP-R, VPU-W)
+ */
+struct vdec_h264_vsi {
+	unsigned char hdr_buf[HDR_PARSING_BUF_SZ];
+	uint64_t pred_buf_dma;
+	uint64_t mv_buf_dma[H264_MAX_FB_NUM];
+	struct h264_ring_fb_list list_free;
+	struct h264_ring_fb_list list_disp;
+	struct vdec_h264_dec_info dec;
+	struct vdec_pic_info pic;
+	struct v4l2_rect crop;
+};
+
+/**
+ * struct vdec_h264_inst - h264 decoder instance
+ * @num_nalu : how many nalus be decoded
+ * @ctx      : point to mtk_vcodec_dec_ctx
+ * @pred_buf : HW working predication buffer
+ * @mv_buf   : HW working motion vector buffer
+ * @vpu      : VPU instance
+ * @vsi      : VPU shared information
+ */
+struct vdec_h264_inst {
+	unsigned int num_nalu;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct mtk_vcodec_mem pred_buf;
+	struct mtk_vcodec_mem mv_buf[H264_MAX_FB_NUM];
+	struct vdec_vpu_inst vpu;
+	struct vdec_h264_vsi *vsi;
+};
+
+static unsigned int get_mv_buf_size(unsigned int width, unsigned int height)
+{
+	return HW_MB_STORE_SZ * (width/MB_UNIT_LEN) * (height/MB_UNIT_LEN);
+}
+
+static int allocate_predication_buf(struct vdec_h264_inst *inst)
+{
+	int err = 0;
+
+	inst->pred_buf.size = BUF_PREDICTION_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, &inst->pred_buf);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "failed to allocate ppl buf");
+		return err;
+	}
+
+	inst->vsi->pred_buf_dma = inst->pred_buf.dma_addr;
+	return 0;
+}
+
+static void free_predication_buf(struct vdec_h264_inst *inst)
+{
+	struct mtk_vcodec_mem *mem = NULL;
+
+	inst->vsi->pred_buf_dma = 0;
+	mem = &inst->pred_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+}
+
+static int alloc_mv_buf(struct vdec_h264_inst *inst, struct vdec_pic_info *pic)
+{
+	int i;
+	int err;
+	struct mtk_vcodec_mem *mem = NULL;
+	unsigned int buf_sz = get_mv_buf_size(pic->buf_w, pic->buf_h);
+
+	for (i = 0; i < H264_MAX_FB_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+		mem->size = buf_sz;
+		err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+		if (err) {
+			mtk_vdec_err(inst->ctx, "failed to allocate mv buf");
+			return err;
+		}
+		inst->vsi->mv_buf_dma[i] = mem->dma_addr;
+	}
+
+	return 0;
+}
+
+static void free_mv_buf(struct vdec_h264_inst *inst)
+{
+	int i;
+	struct mtk_vcodec_mem *mem = NULL;
+
+	for (i = 0; i < H264_MAX_FB_NUM; i++) {
+		inst->vsi->mv_buf_dma[i] = 0;
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+	}
+}
+
+static int check_list_validity(struct vdec_h264_inst *inst, bool disp_list)
+{
+	struct h264_ring_fb_list *list;
+
+	list = disp_list ? &inst->vsi->list_disp : &inst->vsi->list_free;
+
+	if (list->count > H264_MAX_FB_NUM ||
+	    list->read_idx >= H264_MAX_FB_NUM ||
+	    list->write_idx >= H264_MAX_FB_NUM) {
+		mtk_vdec_err(inst->ctx, "%s list err: cnt=%d r_idx=%d w_idx=%d",
+			     disp_list ? "disp" : "free", list->count,
+			     list->read_idx, list->write_idx);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void put_fb_to_free(struct vdec_h264_inst *inst, struct vdec_fb *fb)
+{
+	struct h264_ring_fb_list *list;
+
+	if (fb) {
+		if (check_list_validity(inst, false))
+			return;
+
+		list = &inst->vsi->list_free;
+		if (list->count == H264_MAX_FB_NUM) {
+			mtk_vdec_err(inst->ctx, "[FB] put fb free_list full");
+			return;
+		}
+
+		mtk_vdec_debug(inst->ctx, "[FB] put fb into free_list @(%p, %llx)",
+			       fb->base_y.va, (u64)fb->base_y.dma_addr);
+
+		list->fb_list[list->write_idx].vdec_fb_va = (u64)(uintptr_t)fb;
+		list->write_idx = (list->write_idx == H264_MAX_FB_NUM - 1) ?
+				  0 : list->write_idx + 1;
+		list->count++;
+	}
+}
+
+static void get_pic_info(struct vdec_h264_inst *inst,
+			 struct vdec_pic_info *pic)
+{
+	*pic = inst->vsi->pic;
+	mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)",
+		       pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h);
+	mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", pic->fb_sz[0], pic->fb_sz[1]);
+}
+
+static void get_crop_info(struct vdec_h264_inst *inst, struct v4l2_rect *cr)
+{
+	cr->left = inst->vsi->crop.left;
+	cr->top = inst->vsi->crop.top;
+	cr->width = inst->vsi->crop.width;
+	cr->height = inst->vsi->crop.height;
+
+	mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d", cr->left, cr->top,
+		       cr->width, cr->height);
+}
+
+static void get_dpb_size(struct vdec_h264_inst *inst, unsigned int *dpb_sz)
+{
+	*dpb_sz = inst->vsi->dec.dpb_sz;
+	mtk_vdec_debug(inst->ctx, "sz=%d", *dpb_sz);
+}
+
+static int vdec_h264_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_h264_inst *inst = NULL;
+	int err;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = IPI_VDEC_H264;
+	inst->vpu.ctx = ctx;
+
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_h264 init err=%d", err);
+		goto error_free_inst;
+	}
+
+	inst->vsi = (struct vdec_h264_vsi *)inst->vpu.vsi;
+	err = allocate_predication_buf(inst);
+	if (err)
+		goto error_deinit;
+
+	mtk_vdec_debug(ctx, "H264 Instance >> %p", inst);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+error_deinit:
+	vpu_dec_deinit(&inst->vpu);
+
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static void vdec_h264_deinit(void *h_vdec)
+{
+	struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec;
+
+	vpu_dec_deinit(&inst->vpu);
+	free_predication_buf(inst);
+	free_mv_buf(inst);
+
+	kfree(inst);
+}
+
+static int find_start_code(unsigned char *data, unsigned int data_sz)
+{
+	if (data_sz > 3 && data[0] == 0 && data[1] == 0 && data[2] == 1)
+		return 3;
+
+	if (data_sz > 4 && data[0] == 0 && data[1] == 0 && data[2] == 0 &&
+	    data[3] == 1)
+		return 4;
+
+	return -1;
+}
+
+static int vdec_h264_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+			    struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	int nal_start_idx = 0;
+	int err = 0;
+	unsigned int nal_start;
+	unsigned int nal_type;
+	unsigned char *buf;
+	unsigned int buf_sz;
+	unsigned int data[2];
+	uint64_t vdec_fb_va = (u64)(uintptr_t)fb;
+	uint64_t y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0;
+	uint64_t c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0;
+
+	mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx va=%p",
+		       ++inst->num_nalu, y_fb_dma, c_fb_dma, fb);
+
+	/* bs NULL means flush decoder */
+	if (bs == NULL)
+		return vpu_dec_reset(vpu);
+
+	buf = (unsigned char *)bs->va;
+	buf_sz = bs->size;
+	nal_start_idx = find_start_code(buf, buf_sz);
+	if (nal_start_idx < 0) {
+		mtk_vdec_err(inst->ctx, "invalid nal start code");
+		err = -EIO;
+		goto err_free_fb_out;
+	}
+
+	nal_start = buf[nal_start_idx];
+	nal_type = NAL_TYPE(buf[nal_start_idx]);
+	mtk_vdec_debug(inst->ctx, "\n + NALU[%d] type %d +\n", inst->num_nalu,
+		       nal_type);
+
+	if (nal_type == NAL_H264_PPS) {
+		buf_sz -= nal_start_idx;
+		if (buf_sz > HDR_PARSING_BUF_SZ) {
+			err = -EILSEQ;
+			goto err_free_fb_out;
+		}
+		memcpy(inst->vsi->hdr_buf, buf + nal_start_idx, buf_sz);
+	}
+
+	inst->vsi->dec.bs_dma = (uint64_t)bs->dma_addr;
+	inst->vsi->dec.y_fb_dma = y_fb_dma;
+	inst->vsi->dec.c_fb_dma = c_fb_dma;
+	inst->vsi->dec.vdec_fb_va = vdec_fb_va;
+
+	data[0] = buf_sz;
+	data[1] = nal_start;
+	err = vpu_dec_start(vpu, data, 2);
+	if (err) {
+		if (err > 0 && (DEC_ERR_RET(err) == H264_ERR_NOT_VALID)) {
+			mtk_vdec_err(inst->ctx, "- error bitstream - err = %d -", err);
+			err = -EIO;
+		}
+		goto err_free_fb_out;
+	}
+
+	*res_chg = inst->vsi->dec.resolution_changed;
+	if (*res_chg) {
+		struct vdec_pic_info pic;
+
+		mtk_vdec_debug(inst->ctx, "- resolution changed -");
+		get_pic_info(inst, &pic);
+
+		if (inst->vsi->dec.realloc_mv_buf) {
+			err = alloc_mv_buf(inst, &pic);
+			if (err)
+				goto err_free_fb_out;
+		}
+	}
+
+	if (nal_type == NAL_NON_IDR_SLICE || nal_type == NAL_IDR_SLICE) {
+		/* wait decoder done interrupt */
+		err = mtk_vcodec_wait_for_done_ctx(inst->ctx,
+						   MTK_INST_IRQ_RECEIVED,
+						   WAIT_INTR_TIMEOUT_MS, 0);
+		if (err)
+			goto err_free_fb_out;
+
+		vpu_dec_end(vpu);
+	}
+
+	mtk_vdec_debug(inst->ctx, "\n - NALU[%d] type=%d -\n", inst->num_nalu, nal_type);
+	return 0;
+
+err_free_fb_out:
+	put_fb_to_free(inst, fb);
+	mtk_vdec_err(inst->ctx, "\n - NALU[%d] err=%d -\n", inst->num_nalu, err);
+	return err;
+}
+
+static void vdec_h264_get_fb(struct vdec_h264_inst *inst,
+			     struct h264_ring_fb_list *list,
+			     bool disp_list, struct vdec_fb **out_fb)
+{
+	struct vdec_fb *fb;
+
+	if (check_list_validity(inst, disp_list))
+		return;
+
+	if (list->count == 0) {
+		mtk_vdec_debug(inst->ctx, "[FB] there is no %s fb", disp_list ? "disp" : "free");
+		*out_fb = NULL;
+		return;
+	}
+
+	fb = (struct vdec_fb *)
+		(uintptr_t)list->fb_list[list->read_idx].vdec_fb_va;
+	fb->status |= (disp_list ? FB_ST_DISPLAY : FB_ST_FREE);
+
+	*out_fb = fb;
+	mtk_vdec_debug(inst->ctx, "[FB] get %s fb st=%d poc=%d %llx",
+		       disp_list ? "disp" : "free",
+		       fb->status, list->fb_list[list->read_idx].poc,
+		       list->fb_list[list->read_idx].vdec_fb_va);
+
+	list->read_idx = (list->read_idx == H264_MAX_FB_NUM - 1) ?
+			 0 : list->read_idx + 1;
+	list->count--;
+}
+
+static int vdec_h264_get_param(void *h_vdec, enum vdec_get_param_type type,
+			       void *out)
+{
+	struct vdec_h264_inst *inst = (struct vdec_h264_inst *)h_vdec;
+
+	switch (type) {
+	case GET_PARAM_DISP_FRAME_BUFFER:
+		vdec_h264_get_fb(inst, &inst->vsi->list_disp, true, out);
+		break;
+
+	case GET_PARAM_FREE_FRAME_BUFFER:
+		vdec_h264_get_fb(inst, &inst->vsi->list_free, false, out);
+		break;
+
+	case GET_PARAM_PIC_INFO:
+		get_pic_info(inst, out);
+		break;
+
+	case GET_PARAM_DPB_SIZE:
+		get_dpb_size(inst, out);
+		break;
+
+	case GET_PARAM_CROP_INFO:
+		get_crop_info(inst, out);
+		break;
+
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+const struct vdec_common_if vdec_h264_if = {
+	.init		= vdec_h264_init,
+	.decode		= vdec_h264_decode,
+	.get_param	= vdec_h264_get_param,
+	.deinit		= vdec_h264_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c
new file mode 100644
index 0000000000..5ca20d75dc
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.c
@@ -0,0 +1,322 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2022 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include "vdec_h264_req_common.h"
+
+/* get used parameters for sps/pps */
+#define GET_MTK_VDEC_FLAG(cond, flag) \
+	{ dst_param->cond = ((src_param->flags & flag) ? (1) : (0)); }
+#define GET_MTK_VDEC_PARAM(param) \
+	{ dst_param->param = src_param->param; }
+
+void mtk_vdec_h264_get_ref_list(u8 *ref_list,
+				const struct v4l2_h264_reference *v4l2_ref_list,
+				int num_valid)
+{
+	u32 i;
+
+	/*
+	 * TODO The firmware does not support field decoding. Future
+	 * implementation must use v4l2_ref_list[i].fields to obtain
+	 * the reference field parity.
+	 */
+
+	for (i = 0; i < num_valid; i++)
+		ref_list[i] = v4l2_ref_list[i].index;
+
+	/*
+	 * The firmware expects unused reflist entries to have the value 0x20.
+	 */
+	memset(&ref_list[num_valid], 0x20, 32 - num_valid);
+}
+
+void *mtk_vdec_h264_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id)
+{
+	struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id);
+
+	if (!ctrl)
+		return ERR_PTR(-EINVAL);
+
+	return ctrl->p_cur.p;
+}
+
+void mtk_vdec_h264_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx,
+				 struct slice_api_h264_decode_param *decode_params,
+				 struct mtk_h264_dpb_info *h264_dpb_info)
+{
+	const struct slice_h264_dpb_entry *dpb;
+	struct vb2_queue *vq;
+	struct vb2_buffer *vb;
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	int index;
+
+	vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+
+	for (index = 0; index < V4L2_H264_NUM_DPB_ENTRIES; index++) {
+		dpb = &decode_params->dpb[index];
+		if (!(dpb->flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE)) {
+			h264_dpb_info[index].reference_flag = 0;
+			continue;
+		}
+
+		vb = vb2_find_buffer(vq, dpb->reference_ts);
+		if (!vb) {
+			dev_err(&ctx->dev->plat_dev->dev,
+				"Reference invalid: dpb_index(%d) reference_ts(%lld)",
+				index, dpb->reference_ts);
+			continue;
+		}
+
+		/* 1 for short term reference, 2 for long term reference */
+		if (!(dpb->flags & V4L2_H264_DPB_ENTRY_FLAG_LONG_TERM))
+			h264_dpb_info[index].reference_flag = 1;
+		else
+			h264_dpb_info[index].reference_flag = 2;
+
+		vb2_v4l2 = container_of(vb, struct vb2_v4l2_buffer, vb2_buf);
+		h264_dpb_info[index].field = vb2_v4l2->field;
+
+		h264_dpb_info[index].y_dma_addr =
+			vb2_dma_contig_plane_dma_addr(vb, 0);
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2)
+			h264_dpb_info[index].c_dma_addr =
+				vb2_dma_contig_plane_dma_addr(vb, 1);
+		else
+			h264_dpb_info[index].c_dma_addr =
+				h264_dpb_info[index].y_dma_addr +
+				ctx->picinfo.fb_sz[0];
+	}
+}
+
+void mtk_vdec_h264_copy_sps_params(struct mtk_h264_sps_param *dst_param,
+				   const struct v4l2_ctrl_h264_sps *src_param)
+{
+	GET_MTK_VDEC_PARAM(chroma_format_idc);
+	GET_MTK_VDEC_PARAM(bit_depth_luma_minus8);
+	GET_MTK_VDEC_PARAM(bit_depth_chroma_minus8);
+	GET_MTK_VDEC_PARAM(log2_max_frame_num_minus4);
+	GET_MTK_VDEC_PARAM(pic_order_cnt_type);
+	GET_MTK_VDEC_PARAM(log2_max_pic_order_cnt_lsb_minus4);
+	GET_MTK_VDEC_PARAM(max_num_ref_frames);
+	GET_MTK_VDEC_PARAM(pic_width_in_mbs_minus1);
+	GET_MTK_VDEC_PARAM(pic_height_in_map_units_minus1);
+
+	GET_MTK_VDEC_FLAG(separate_colour_plane_flag,
+			  V4L2_H264_SPS_FLAG_SEPARATE_COLOUR_PLANE);
+	GET_MTK_VDEC_FLAG(qpprime_y_zero_transform_bypass_flag,
+			  V4L2_H264_SPS_FLAG_QPPRIME_Y_ZERO_TRANSFORM_BYPASS);
+	GET_MTK_VDEC_FLAG(delta_pic_order_always_zero_flag,
+			  V4L2_H264_SPS_FLAG_DELTA_PIC_ORDER_ALWAYS_ZERO);
+	GET_MTK_VDEC_FLAG(frame_mbs_only_flag,
+			  V4L2_H264_SPS_FLAG_FRAME_MBS_ONLY);
+	GET_MTK_VDEC_FLAG(mb_adaptive_frame_field_flag,
+			  V4L2_H264_SPS_FLAG_MB_ADAPTIVE_FRAME_FIELD);
+	GET_MTK_VDEC_FLAG(direct_8x8_inference_flag,
+			  V4L2_H264_SPS_FLAG_DIRECT_8X8_INFERENCE);
+}
+
+void mtk_vdec_h264_copy_pps_params(struct mtk_h264_pps_param *dst_param,
+				   const struct v4l2_ctrl_h264_pps *src_param)
+{
+	GET_MTK_VDEC_PARAM(num_ref_idx_l0_default_active_minus1);
+	GET_MTK_VDEC_PARAM(num_ref_idx_l1_default_active_minus1);
+	GET_MTK_VDEC_PARAM(weighted_bipred_idc);
+	GET_MTK_VDEC_PARAM(pic_init_qp_minus26);
+	GET_MTK_VDEC_PARAM(chroma_qp_index_offset);
+	GET_MTK_VDEC_PARAM(second_chroma_qp_index_offset);
+
+	GET_MTK_VDEC_FLAG(entropy_coding_mode_flag,
+			  V4L2_H264_PPS_FLAG_ENTROPY_CODING_MODE);
+	GET_MTK_VDEC_FLAG(pic_order_present_flag,
+			  V4L2_H264_PPS_FLAG_BOTTOM_FIELD_PIC_ORDER_IN_FRAME_PRESENT);
+	GET_MTK_VDEC_FLAG(weighted_pred_flag,
+			  V4L2_H264_PPS_FLAG_WEIGHTED_PRED);
+	GET_MTK_VDEC_FLAG(deblocking_filter_control_present_flag,
+			  V4L2_H264_PPS_FLAG_DEBLOCKING_FILTER_CONTROL_PRESENT);
+	GET_MTK_VDEC_FLAG(constrained_intra_pred_flag,
+			  V4L2_H264_PPS_FLAG_CONSTRAINED_INTRA_PRED);
+	GET_MTK_VDEC_FLAG(redundant_pic_cnt_present_flag,
+			  V4L2_H264_PPS_FLAG_REDUNDANT_PIC_CNT_PRESENT);
+	GET_MTK_VDEC_FLAG(transform_8x8_mode_flag,
+			  V4L2_H264_PPS_FLAG_TRANSFORM_8X8_MODE);
+	GET_MTK_VDEC_FLAG(scaling_matrix_present_flag,
+			  V4L2_H264_PPS_FLAG_SCALING_MATRIX_PRESENT);
+}
+
+void mtk_vdec_h264_copy_slice_hd_params(struct mtk_h264_slice_hd_param *dst_param,
+					const struct v4l2_ctrl_h264_slice_params *src_param,
+					const struct v4l2_ctrl_h264_decode_params *dec_param)
+{
+	int temp;
+
+	GET_MTK_VDEC_PARAM(first_mb_in_slice);
+	GET_MTK_VDEC_PARAM(slice_type);
+	GET_MTK_VDEC_PARAM(cabac_init_idc);
+	GET_MTK_VDEC_PARAM(slice_qp_delta);
+	GET_MTK_VDEC_PARAM(disable_deblocking_filter_idc);
+	GET_MTK_VDEC_PARAM(slice_alpha_c0_offset_div2);
+	GET_MTK_VDEC_PARAM(slice_beta_offset_div2);
+	GET_MTK_VDEC_PARAM(num_ref_idx_l0_active_minus1);
+	GET_MTK_VDEC_PARAM(num_ref_idx_l1_active_minus1);
+
+	dst_param->frame_num = dec_param->frame_num;
+	dst_param->pic_order_cnt_lsb = dec_param->pic_order_cnt_lsb;
+
+	dst_param->delta_pic_order_cnt_bottom =
+		dec_param->delta_pic_order_cnt_bottom;
+	dst_param->delta_pic_order_cnt0 =
+		dec_param->delta_pic_order_cnt0;
+	dst_param->delta_pic_order_cnt1 =
+		dec_param->delta_pic_order_cnt1;
+
+	temp = dec_param->flags & V4L2_H264_DECODE_PARAM_FLAG_FIELD_PIC;
+	dst_param->field_pic_flag = temp ? 1 : 0;
+
+	temp = dec_param->flags & V4L2_H264_DECODE_PARAM_FLAG_BOTTOM_FIELD;
+	dst_param->bottom_field_flag = temp ? 1 : 0;
+
+	GET_MTK_VDEC_FLAG(direct_spatial_mv_pred_flag,
+			  V4L2_H264_SLICE_FLAG_DIRECT_SPATIAL_MV_PRED);
+}
+
+void mtk_vdec_h264_copy_scaling_matrix(struct slice_api_h264_scaling_matrix *dst_matrix,
+				       const struct v4l2_ctrl_h264_scaling_matrix *src_matrix)
+{
+	memcpy(dst_matrix->scaling_list_4x4, src_matrix->scaling_list_4x4,
+	       sizeof(dst_matrix->scaling_list_4x4));
+
+	memcpy(dst_matrix->scaling_list_8x8, src_matrix->scaling_list_8x8,
+	       sizeof(dst_matrix->scaling_list_8x8));
+}
+
+void
+mtk_vdec_h264_copy_decode_params(struct slice_api_h264_decode_param *dst_params,
+				 const struct v4l2_ctrl_h264_decode_params *src_params,
+				 const struct v4l2_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES])
+{
+	struct slice_h264_dpb_entry *dst_entry;
+	const struct v4l2_h264_dpb_entry *src_entry;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(dst_params->dpb); i++) {
+		dst_entry = &dst_params->dpb[i];
+		src_entry = &dpb[i];
+
+		dst_entry->reference_ts = src_entry->reference_ts;
+		dst_entry->frame_num = src_entry->frame_num;
+		dst_entry->pic_num = src_entry->pic_num;
+		dst_entry->top_field_order_cnt = src_entry->top_field_order_cnt;
+		dst_entry->bottom_field_order_cnt =
+			src_entry->bottom_field_order_cnt;
+		dst_entry->flags = src_entry->flags;
+	}
+
+	/* num_slices is a leftover from the old H.264 support and is ignored
+	 * by the firmware.
+	 */
+	dst_params->num_slices = 0;
+	dst_params->nal_ref_idc = src_params->nal_ref_idc;
+	dst_params->top_field_order_cnt = src_params->top_field_order_cnt;
+	dst_params->bottom_field_order_cnt = src_params->bottom_field_order_cnt;
+	dst_params->flags = src_params->flags;
+}
+
+static bool mtk_vdec_h264_dpb_entry_match(const struct v4l2_h264_dpb_entry *a,
+					  const struct v4l2_h264_dpb_entry *b)
+{
+	return a->top_field_order_cnt == b->top_field_order_cnt &&
+	       a->bottom_field_order_cnt == b->bottom_field_order_cnt;
+}
+
+/*
+ * Move DPB entries of dec_param that refer to a frame already existing in dpb
+ * into the already existing slot in dpb, and move other entries into new slots.
+ *
+ * This function is an adaptation of the similarly-named function in
+ * hantro_h264.c.
+ */
+void mtk_vdec_h264_update_dpb(const struct v4l2_ctrl_h264_decode_params *dec_param,
+			      struct v4l2_h264_dpb_entry *dpb)
+{
+	DECLARE_BITMAP(new, ARRAY_SIZE(dec_param->dpb)) = { 0, };
+	DECLARE_BITMAP(in_use, ARRAY_SIZE(dec_param->dpb)) = { 0, };
+	DECLARE_BITMAP(used, ARRAY_SIZE(dec_param->dpb)) = { 0, };
+	unsigned int i, j;
+
+	/* Disable all entries by default, and mark the ones in use. */
+	for (i = 0; i < ARRAY_SIZE(dec_param->dpb); i++) {
+		if (dpb[i].flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE)
+			set_bit(i, in_use);
+		dpb[i].flags &= ~V4L2_H264_DPB_ENTRY_FLAG_ACTIVE;
+	}
+
+	/* Try to match new DPB entries with existing ones by their POCs. */
+	for (i = 0; i < ARRAY_SIZE(dec_param->dpb); i++) {
+		const struct v4l2_h264_dpb_entry *ndpb = &dec_param->dpb[i];
+
+		if (!(ndpb->flags & V4L2_H264_DPB_ENTRY_FLAG_ACTIVE))
+			continue;
+
+		/*
+		 * To cut off some comparisons, iterate only on target DPB
+		 * entries were already used.
+		 */
+		for_each_set_bit(j, in_use, ARRAY_SIZE(dec_param->dpb)) {
+			struct v4l2_h264_dpb_entry *cdpb;
+
+			cdpb = &dpb[j];
+			if (!mtk_vdec_h264_dpb_entry_match(cdpb, ndpb))
+				continue;
+
+			*cdpb = *ndpb;
+			set_bit(j, used);
+			/* Don't reiterate on this one. */
+			clear_bit(j, in_use);
+			break;
+		}
+
+		if (j == ARRAY_SIZE(dec_param->dpb))
+			set_bit(i, new);
+	}
+
+	/* For entries that could not be matched, use remaining free slots. */
+	for_each_set_bit(i, new, ARRAY_SIZE(dec_param->dpb)) {
+		const struct v4l2_h264_dpb_entry *ndpb = &dec_param->dpb[i];
+		struct v4l2_h264_dpb_entry *cdpb;
+
+		/*
+		 * Both arrays are of the same sizes, so there is no way
+		 * we can end up with no space in target array, unless
+		 * something is buggy.
+		 */
+		j = find_first_zero_bit(used, ARRAY_SIZE(dec_param->dpb));
+		if (WARN_ON(j >= ARRAY_SIZE(dec_param->dpb)))
+			return;
+
+		cdpb = &dpb[j];
+		*cdpb = *ndpb;
+		set_bit(j, used);
+	}
+}
+
+unsigned int mtk_vdec_h264_get_mv_buf_size(unsigned int width, unsigned int height)
+{
+	int unit_size = (width / MB_UNIT_LEN) * (height / MB_UNIT_LEN) + 8;
+
+	return HW_MB_STORE_SZ * unit_size;
+}
+
+int mtk_vdec_h264_find_start_code(unsigned char *data, unsigned int data_sz)
+{
+	if (data_sz > 3 && data[0] == 0 && data[1] == 0 && data[2] == 1)
+		return 3;
+
+	if (data_sz > 4 && data[0] == 0 && data[1] == 0 && data[2] == 0 &&
+	    data[3] == 1)
+		return 4;
+
+	return -1;
+}
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h
new file mode 100644
index 0000000000..ac82be3360
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_common.h
@@ -0,0 +1,277 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2022 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#ifndef _VDEC_H264_REQ_COMMON_H_
+#define _VDEC_H264_REQ_COMMON_H_
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/v4l2-h264.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "../mtk_vcodec_dec_drv.h"
+
+#define NAL_NON_IDR_SLICE			0x01
+#define NAL_IDR_SLICE				0x05
+#define NAL_TYPE(value)				((value) & 0x1F)
+
+#define BUF_PREDICTION_SZ			(64 * 4096)
+#define MB_UNIT_LEN				16
+
+/* motion vector size (bytes) for every macro block */
+#define HW_MB_STORE_SZ				64
+
+#define H264_MAX_MV_NUM				32
+
+/**
+ * struct mtk_h264_dpb_info  - h264 dpb information
+ *
+ * @y_dma_addr:	Y bitstream physical address
+ * @c_dma_addr:	CbCr bitstream physical address
+ * @reference_flag:	reference picture flag (short/long term reference picture)
+ * @field:		field picture flag
+ */
+struct mtk_h264_dpb_info {
+	dma_addr_t y_dma_addr;
+	dma_addr_t c_dma_addr;
+	int reference_flag;
+	int field;
+};
+
+/*
+ * struct mtk_h264_sps_param  - parameters for sps
+ */
+struct mtk_h264_sps_param {
+	unsigned char chroma_format_idc;
+	unsigned char bit_depth_luma_minus8;
+	unsigned char bit_depth_chroma_minus8;
+	unsigned char log2_max_frame_num_minus4;
+	unsigned char pic_order_cnt_type;
+	unsigned char log2_max_pic_order_cnt_lsb_minus4;
+	unsigned char max_num_ref_frames;
+	unsigned char separate_colour_plane_flag;
+	unsigned short pic_width_in_mbs_minus1;
+	unsigned short pic_height_in_map_units_minus1;
+	unsigned int max_frame_nums;
+	unsigned char qpprime_y_zero_transform_bypass_flag;
+	unsigned char delta_pic_order_always_zero_flag;
+	unsigned char frame_mbs_only_flag;
+	unsigned char mb_adaptive_frame_field_flag;
+	unsigned char direct_8x8_inference_flag;
+	unsigned char reserved[3];
+};
+
+/*
+ * struct mtk_h264_pps_param  - parameters for pps
+ */
+struct mtk_h264_pps_param {
+	unsigned char num_ref_idx_l0_default_active_minus1;
+	unsigned char num_ref_idx_l1_default_active_minus1;
+	unsigned char weighted_bipred_idc;
+	char pic_init_qp_minus26;
+	char chroma_qp_index_offset;
+	char second_chroma_qp_index_offset;
+	unsigned char entropy_coding_mode_flag;
+	unsigned char pic_order_present_flag;
+	unsigned char deblocking_filter_control_present_flag;
+	unsigned char constrained_intra_pred_flag;
+	unsigned char weighted_pred_flag;
+	unsigned char redundant_pic_cnt_present_flag;
+	unsigned char transform_8x8_mode_flag;
+	unsigned char scaling_matrix_present_flag;
+	unsigned char reserved[2];
+};
+
+/*
+ * struct mtk_h264_slice_hd_param  - parameters for slice header
+ */
+struct mtk_h264_slice_hd_param {
+	unsigned int first_mb_in_slice;
+	unsigned int field_pic_flag;
+	unsigned int slice_type;
+	unsigned int frame_num;
+	int pic_order_cnt_lsb;
+	int delta_pic_order_cnt_bottom;
+	unsigned int bottom_field_flag;
+	unsigned int direct_spatial_mv_pred_flag;
+	int delta_pic_order_cnt0;
+	int delta_pic_order_cnt1;
+	unsigned int cabac_init_idc;
+	int slice_qp_delta;
+	unsigned int disable_deblocking_filter_idc;
+	int slice_alpha_c0_offset_div2;
+	int slice_beta_offset_div2;
+	unsigned int num_ref_idx_l0_active_minus1;
+	unsigned int num_ref_idx_l1_active_minus1;
+	unsigned int reserved;
+};
+
+/*
+ * struct slice_api_h264_scaling_matrix  - parameters for scaling list
+ */
+struct slice_api_h264_scaling_matrix {
+	unsigned char scaling_list_4x4[6][16];
+	unsigned char scaling_list_8x8[6][64];
+};
+
+/*
+ * struct slice_h264_dpb_entry  - each dpb information
+ */
+struct slice_h264_dpb_entry {
+	unsigned long long reference_ts;
+	unsigned short frame_num;
+	unsigned short pic_num;
+	/* Note that field is indicated by v4l2_buffer.field */
+	int top_field_order_cnt;
+	int bottom_field_order_cnt;
+	unsigned int flags;
+};
+
+/*
+ * struct slice_api_h264_decode_param - parameters for decode.
+ */
+struct slice_api_h264_decode_param {
+	struct slice_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES];
+	unsigned short num_slices;
+	unsigned short nal_ref_idc;
+	unsigned char ref_pic_list_p0[32];
+	unsigned char ref_pic_list_b0[32];
+	unsigned char ref_pic_list_b1[32];
+	int top_field_order_cnt;
+	int bottom_field_order_cnt;
+	unsigned int flags;
+};
+
+/**
+ * struct h264_fb - h264 decode frame buffer information
+ *
+ * @vdec_fb_va:	virtual address of struct vdec_fb
+ * @y_fb_dma:		dma address of Y frame buffer (luma)
+ * @c_fb_dma:		dma address of C frame buffer (chroma)
+ * @poc:		picture order count of frame buffer
+ * @reserved:		for 8 bytes alignment
+ */
+struct h264_fb {
+	u64 vdec_fb_va;
+	u64 y_fb_dma;
+	u64 c_fb_dma;
+	s32 poc;
+	u32 reserved;
+};
+
+/**
+ * mtk_vdec_h264_get_ref_list - translate V4L2 reference list
+ *
+ * @ref_list:		Mediatek reference picture list
+ * @v4l2_ref_list:	V4L2 reference picture list
+ * @num_valid:		used reference number
+ */
+void mtk_vdec_h264_get_ref_list(u8 *ref_list,
+				const struct v4l2_h264_reference *v4l2_ref_list,
+				int num_valid);
+
+/**
+ * mtk_vdec_h264_get_ctrl_ptr - get each CID contrl address.
+ *
+ * @ctx:	v4l2 ctx
+ * @id:	CID control ID
+ *
+ * Return: returns CID ctrl address.
+ */
+void *mtk_vdec_h264_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id);
+
+/**
+ * mtk_vdec_h264_fill_dpb_info - get each CID contrl address.
+ *
+ * @ctx:		v4l2 ctx
+ * @decode_params:	slice decode params
+ * @h264_dpb_info:	dpb buffer information
+ */
+void mtk_vdec_h264_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx,
+				 struct slice_api_h264_decode_param *decode_params,
+				 struct mtk_h264_dpb_info *h264_dpb_info);
+
+/**
+ * mtk_vdec_h264_copy_sps_params - get sps params.
+ *
+ * @dst_param:	sps params for hw decoder
+ * @src_param:	sps params from user driver
+ */
+void mtk_vdec_h264_copy_sps_params(struct mtk_h264_sps_param *dst_param,
+				   const struct v4l2_ctrl_h264_sps *src_param);
+
+/**
+ * mtk_vdec_h264_copy_pps_params - get pps params.
+ *
+ * @dst_param:	pps params for hw decoder
+ * @src_param:	pps params from user driver
+ */
+void mtk_vdec_h264_copy_pps_params(struct mtk_h264_pps_param *dst_param,
+				   const struct v4l2_ctrl_h264_pps *src_param);
+
+/**
+ * mtk_vdec_h264_copy_slice_hd_params - get slice header params.
+ *
+ * @dst_param:	slice params for hw decoder
+ * @src_param:	slice params from user driver
+ * @dec_param:	decode params from user driver
+ */
+void mtk_vdec_h264_copy_slice_hd_params(struct mtk_h264_slice_hd_param *dst_param,
+					const struct v4l2_ctrl_h264_slice_params *src_param,
+					const struct v4l2_ctrl_h264_decode_params *dec_param);
+
+/**
+ * mtk_vdec_h264_copy_scaling_matrix - get each CID contrl address.
+ *
+ * @dst_matrix:	scaling list params for hw decoder
+ * @src_matrix:	scaling list params from user driver
+ */
+void mtk_vdec_h264_copy_scaling_matrix(struct slice_api_h264_scaling_matrix *dst_matrix,
+				       const struct v4l2_ctrl_h264_scaling_matrix *src_matrix);
+
+/**
+ * mtk_vdec_h264_copy_decode_params - get decode params.
+ *
+ * @dst_params:	dst params for hw decoder
+ * @src_params:	decode params from user driver
+ * @dpb:		dpb information
+ */
+void
+mtk_vdec_h264_copy_decode_params(struct slice_api_h264_decode_param *dst_params,
+				 const struct v4l2_ctrl_h264_decode_params *src_params,
+				 const struct v4l2_h264_dpb_entry dpb[V4L2_H264_NUM_DPB_ENTRIES]);
+
+/**
+ * mtk_vdec_h264_update_dpb - updata dpb list.
+ *
+ * @dec_param:	v4l2 control decode params
+ * @dpb:	dpb entry informaton
+ */
+void mtk_vdec_h264_update_dpb(const struct v4l2_ctrl_h264_decode_params *dec_param,
+			      struct v4l2_h264_dpb_entry *dpb);
+
+/**
+ * mtk_vdec_h264_find_start_code - find h264 start code using sofeware.
+ *
+ * @data:	input buffer address
+ * @data_sz:	input buffer size
+ *
+ * Return: returns start code position.
+ */
+int mtk_vdec_h264_find_start_code(unsigned char *data, unsigned int data_sz);
+
+/**
+ * mtk_vdec_h264_get_mv_buf_size - get mv buffer size.
+ *
+ * @width:	picture width
+ * @height:	picture height
+ *
+ * Return: returns mv buffer size.
+ */
+unsigned int mtk_vdec_h264_get_mv_buf_size(unsigned int width, unsigned int height);
+
+#endif
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c
new file mode 100644
index 0000000000..5600f1df65
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_if.c
@@ -0,0 +1,444 @@
+// SPDX-License-Identifier: GPL-2.0
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/v4l2-h264.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+#include "vdec_h264_req_common.h"
+
+/*
+ * struct mtk_h264_dec_slice_param  - parameters for decode current frame
+ */
+struct mtk_h264_dec_slice_param {
+	struct mtk_h264_sps_param			sps;
+	struct mtk_h264_pps_param			pps;
+	struct slice_api_h264_scaling_matrix		scaling_matrix;
+	struct slice_api_h264_decode_param		decode_params;
+	struct mtk_h264_dpb_info h264_dpb_info[16];
+};
+
+/**
+ * struct vdec_h264_dec_info - decode information
+ * @dpb_sz		: decoding picture buffer size
+ * @resolution_changed  : resoltion change happen
+ * @realloc_mv_buf	: flag to notify driver to re-allocate mv buffer
+ * @cap_num_planes	: number planes of capture buffer
+ * @bs_dma		: Input bit-stream buffer dma address
+ * @y_fb_dma		: Y frame buffer dma address
+ * @c_fb_dma		: C frame buffer dma address
+ * @vdec_fb_va		: VDEC frame buffer struct virtual address
+ */
+struct vdec_h264_dec_info {
+	u32 dpb_sz;
+	u32 resolution_changed;
+	u32 realloc_mv_buf;
+	u32 cap_num_planes;
+	u64 bs_dma;
+	u64 y_fb_dma;
+	u64 c_fb_dma;
+	u64 vdec_fb_va;
+};
+
+/**
+ * struct vdec_h264_vsi - shared memory for decode information exchange
+ *                        between VPU and Host.
+ *                        The memory is allocated by VPU then mapping to Host
+ *                        in vpu_dec_init() and freed in vpu_dec_deinit()
+ *                        by VPU.
+ *                        AP-W/R : AP is writer/reader on this item
+ *                        VPU-W/R: VPU is write/reader on this item
+ * @pred_buf_dma : HW working predication buffer dma address (AP-W, VPU-R)
+ * @mv_buf_dma   : HW working motion vector buffer dma address (AP-W, VPU-R)
+ * @dec          : decode information (AP-R, VPU-W)
+ * @pic          : picture information (AP-R, VPU-W)
+ * @crop         : crop information (AP-R, VPU-W)
+ * @h264_slice_params : the parameters that hardware use to decode
+ */
+struct vdec_h264_vsi {
+	u64 pred_buf_dma;
+	u64 mv_buf_dma[H264_MAX_MV_NUM];
+	struct vdec_h264_dec_info dec;
+	struct vdec_pic_info pic;
+	struct v4l2_rect crop;
+	struct mtk_h264_dec_slice_param h264_slice_params;
+};
+
+/**
+ * struct vdec_h264_slice_inst - h264 decoder instance
+ * @num_nalu : how many nalus be decoded
+ * @ctx      : point to mtk_vcodec_dec_ctx
+ * @pred_buf : HW working predication buffer
+ * @mv_buf   : HW working motion vector buffer
+ * @vpu      : VPU instance
+ * @vsi_ctx  : Local VSI data for this decoding context
+ * @h264_slice_param : the parameters that hardware use to decode
+ * @dpb : decoded picture buffer used to store reference buffer information
+ */
+struct vdec_h264_slice_inst {
+	unsigned int num_nalu;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct mtk_vcodec_mem pred_buf;
+	struct mtk_vcodec_mem mv_buf[H264_MAX_MV_NUM];
+	struct vdec_vpu_inst vpu;
+	struct vdec_h264_vsi vsi_ctx;
+	struct mtk_h264_dec_slice_param h264_slice_param;
+
+	struct v4l2_h264_dpb_entry dpb[16];
+};
+
+static int get_vdec_decode_parameters(struct vdec_h264_slice_inst *inst)
+{
+	const struct v4l2_ctrl_h264_decode_params *dec_params;
+	const struct v4l2_ctrl_h264_sps *sps;
+	const struct v4l2_ctrl_h264_pps *pps;
+	const struct v4l2_ctrl_h264_scaling_matrix *scaling_matrix;
+	struct mtk_h264_dec_slice_param *slice_param = &inst->h264_slice_param;
+	struct v4l2_h264_reflist_builder reflist_builder;
+	struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN];
+	u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0;
+	u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0;
+	u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1;
+
+	dec_params =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS);
+	if (IS_ERR(dec_params))
+		return PTR_ERR(dec_params);
+
+	sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS);
+	if (IS_ERR(sps))
+		return PTR_ERR(sps);
+
+	pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS);
+	if (IS_ERR(pps))
+		return PTR_ERR(pps);
+
+	scaling_matrix =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX);
+	if (IS_ERR(scaling_matrix))
+		return PTR_ERR(scaling_matrix);
+
+	mtk_vdec_h264_update_dpb(dec_params, inst->dpb);
+
+	mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps);
+	mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps);
+	mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, scaling_matrix);
+	mtk_vdec_h264_copy_decode_params(&slice_param->decode_params,
+					 dec_params, inst->dpb);
+	mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params,
+				    slice_param->h264_dpb_info);
+
+	/* Build the reference lists */
+	v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps,
+				       inst->dpb);
+	v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist);
+	v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist,
+				    v4l2_b1_reflist);
+
+	/* Adapt the built lists to the firmware's expectations */
+	mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid);
+
+	memcpy(&inst->vsi_ctx.h264_slice_params, slice_param,
+	       sizeof(inst->vsi_ctx.h264_slice_params));
+
+	return 0;
+}
+
+static int allocate_predication_buf(struct vdec_h264_slice_inst *inst)
+{
+	int err;
+
+	inst->pred_buf.size = BUF_PREDICTION_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, &inst->pred_buf);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "failed to allocate ppl buf");
+		return err;
+	}
+
+	inst->vsi_ctx.pred_buf_dma = inst->pred_buf.dma_addr;
+	return 0;
+}
+
+static void free_predication_buf(struct vdec_h264_slice_inst *inst)
+{
+	struct mtk_vcodec_mem *mem = &inst->pred_buf;
+
+	inst->vsi_ctx.pred_buf_dma = 0;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+}
+
+static int alloc_mv_buf(struct vdec_h264_slice_inst *inst,
+			struct vdec_pic_info *pic)
+{
+	int i;
+	int err;
+	struct mtk_vcodec_mem *mem = NULL;
+	unsigned int buf_sz = mtk_vdec_h264_get_mv_buf_size(pic->buf_w, pic->buf_h);
+
+	mtk_v4l2_vdec_dbg(3, inst->ctx, "size = 0x%x", buf_sz);
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+		mem->size = buf_sz;
+		err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+		if (err) {
+			mtk_vdec_err(inst->ctx, "failed to allocate mv buf");
+			return err;
+		}
+		inst->vsi_ctx.mv_buf_dma[i] = mem->dma_addr;
+	}
+
+	return 0;
+}
+
+static void free_mv_buf(struct vdec_h264_slice_inst *inst)
+{
+	int i;
+	struct mtk_vcodec_mem *mem;
+
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		inst->vsi_ctx.mv_buf_dma[i] = 0;
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+	}
+}
+
+static void get_pic_info(struct vdec_h264_slice_inst *inst,
+			 struct vdec_pic_info *pic)
+{
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.fb_sz[0] = ctx->picinfo.buf_w * ctx->picinfo.buf_h;
+	ctx->picinfo.fb_sz[1] = ctx->picinfo.fb_sz[0] >> 1;
+	inst->vsi_ctx.dec.cap_num_planes =
+		ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes;
+
+	*pic = ctx->picinfo;
+	mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)",
+		       ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+		       ctx->picinfo.buf_w, ctx->picinfo.buf_h);
+	mtk_vdec_debug(inst->ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0],
+		       ctx->picinfo.fb_sz[1]);
+
+	if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w ||
+	    ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) {
+		inst->vsi_ctx.dec.resolution_changed = true;
+		if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w ||
+		    ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h)
+			inst->vsi_ctx.dec.realloc_mv_buf = true;
+
+		mtk_v4l2_vdec_dbg(1, inst->ctx, "ResChg: (%d %d) : old(%d, %d) -> new(%d, %d)",
+				  inst->vsi_ctx.dec.resolution_changed,
+				  inst->vsi_ctx.dec.realloc_mv_buf,
+				  ctx->last_decoded_picinfo.pic_w,
+				  ctx->last_decoded_picinfo.pic_h,
+				  ctx->picinfo.pic_w, ctx->picinfo.pic_h);
+	}
+}
+
+static void get_crop_info(struct vdec_h264_slice_inst *inst, struct v4l2_rect *cr)
+{
+	cr->left = inst->vsi_ctx.crop.left;
+	cr->top = inst->vsi_ctx.crop.top;
+	cr->width = inst->vsi_ctx.crop.width;
+	cr->height = inst->vsi_ctx.crop.height;
+
+	mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static void get_dpb_size(struct vdec_h264_slice_inst *inst, unsigned int *dpb_sz)
+{
+	*dpb_sz = inst->vsi_ctx.dec.dpb_sz;
+	mtk_vdec_debug(inst->ctx, "sz=%d", *dpb_sz);
+}
+
+static int vdec_h264_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_h264_slice_inst *inst;
+	int err;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = SCP_IPI_VDEC_H264;
+	inst->vpu.ctx = ctx;
+
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_h264 init err=%d", err);
+		goto error_free_inst;
+	}
+
+	memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx));
+	inst->vsi_ctx.dec.resolution_changed = true;
+	inst->vsi_ctx.dec.realloc_mv_buf = true;
+
+	err = allocate_predication_buf(inst);
+	if (err)
+		goto error_deinit;
+
+	mtk_vdec_debug(ctx, "struct size = %zu,%zu,%zu,%zu\n",
+		       sizeof(struct mtk_h264_sps_param),
+		       sizeof(struct mtk_h264_pps_param),
+		       sizeof(struct mtk_h264_dec_slice_param),
+		       sizeof(struct mtk_h264_dpb_info));
+
+	mtk_vdec_debug(ctx, "H264 Instance >> %p", inst);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+error_deinit:
+	vpu_dec_deinit(&inst->vpu);
+
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static void vdec_h264_slice_deinit(void *h_vdec)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+
+	vpu_dec_deinit(&inst->vpu);
+	free_predication_buf(inst);
+	free_mv_buf(inst);
+
+	kfree(inst);
+}
+
+static int vdec_h264_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				  struct vdec_fb *unused, bool *res_chg)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+	const struct v4l2_ctrl_h264_decode_params *dec_params =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS);
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	struct mtk_video_dec_buf *src_buf_info;
+	struct mtk_video_dec_buf *dst_buf_info;
+	struct vdec_fb *fb;
+	u32 data[2];
+	u64 y_fb_dma;
+	u64 c_fb_dma;
+	int err;
+
+	inst->num_nalu++;
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vpu_dec_reset(vpu);
+
+	fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx);
+	src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer);
+	dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer);
+
+	y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0;
+	c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0;
+
+	mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx va=%p",
+		       inst->num_nalu, y_fb_dma, c_fb_dma, fb);
+
+	inst->vsi_ctx.dec.bs_dma = (uint64_t)bs->dma_addr;
+	inst->vsi_ctx.dec.y_fb_dma = y_fb_dma;
+	inst->vsi_ctx.dec.c_fb_dma = c_fb_dma;
+	inst->vsi_ctx.dec.vdec_fb_va = (u64)(uintptr_t)fb;
+
+	v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb,
+				   &dst_buf_info->m2m_buf.vb, true);
+	err = get_vdec_decode_parameters(inst);
+	if (err)
+		goto err_free_fb_out;
+
+	data[0] = bs->size;
+	/*
+	 * Reconstruct the first byte of the NAL unit, as the firmware requests
+	 * that information to be passed even though it is present in the stream
+	 * itself...
+	 */
+	data[1] = (dec_params->nal_ref_idc << 5) |
+		  ((dec_params->flags & V4L2_H264_DECODE_PARAM_FLAG_IDR_PIC)
+			? 0x5 : 0x1);
+
+	*res_chg = inst->vsi_ctx.dec.resolution_changed;
+	if (*res_chg) {
+		mtk_vdec_debug(inst->ctx, "- resolution changed -");
+		if (inst->vsi_ctx.dec.realloc_mv_buf) {
+			err = alloc_mv_buf(inst, &inst->ctx->picinfo);
+			inst->vsi_ctx.dec.realloc_mv_buf = false;
+			if (err)
+				goto err_free_fb_out;
+		}
+		*res_chg = false;
+	}
+
+	memcpy(inst->vpu.vsi, &inst->vsi_ctx, sizeof(inst->vsi_ctx));
+	err = vpu_dec_start(vpu, data, 2);
+	if (err)
+		goto err_free_fb_out;
+
+	/* wait decoder done interrupt */
+	err = mtk_vcodec_wait_for_done_ctx(inst->ctx,
+					   MTK_INST_IRQ_RECEIVED,
+					   WAIT_INTR_TIMEOUT_MS, 0);
+	if (err)
+		goto err_free_fb_out;
+	vpu_dec_end(vpu);
+
+	memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx));
+	mtk_vdec_debug(inst->ctx, "\n - NALU[%d]", inst->num_nalu);
+	return 0;
+
+err_free_fb_out:
+	mtk_vdec_err(inst->ctx, "\n - NALU[%d] err=%d -\n", inst->num_nalu, err);
+	return err;
+}
+
+static int vdec_h264_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		get_pic_info(inst, out);
+		break;
+
+	case GET_PARAM_DPB_SIZE:
+		get_dpb_size(inst, out);
+		break;
+
+	case GET_PARAM_CROP_INFO:
+		get_crop_info(inst, out);
+		break;
+
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+const struct vdec_common_if vdec_h264_slice_if = {
+	.init		= vdec_h264_slice_init,
+	.decode		= vdec_h264_slice_decode,
+	.get_param	= vdec_h264_slice_get_param,
+	.deinit		= vdec_h264_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c
new file mode 100644
index 0000000000..0e741e0dc8
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_h264_req_multi_if.c
@@ -0,0 +1,848 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2022 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/v4l2-h264.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+#include "vdec_h264_req_common.h"
+
+/**
+ * enum vdec_h264_core_dec_err_type  - core decode error type
+ *
+ * @TRANS_BUFFER_FULL: trans buffer is full
+ * @SLICE_HEADER_FULL: slice header buffer is full
+ */
+enum vdec_h264_core_dec_err_type {
+	TRANS_BUFFER_FULL = 1,
+	SLICE_HEADER_FULL,
+};
+
+/**
+ * struct vdec_h264_slice_lat_dec_param  - parameters for decode current frame
+ *
+ * @sps:		h264 sps syntax parameters
+ * @pps:		h264 pps syntax parameters
+ * @slice_header:	h264 slice header syntax parameters
+ * @scaling_matrix:	h264 scaling list parameters
+ * @decode_params:	decoder parameters of each frame used for hardware decode
+ * @h264_dpb_info:	dpb reference list
+ */
+struct vdec_h264_slice_lat_dec_param {
+	struct mtk_h264_sps_param sps;
+	struct mtk_h264_pps_param pps;
+	struct mtk_h264_slice_hd_param slice_header;
+	struct slice_api_h264_scaling_matrix scaling_matrix;
+	struct slice_api_h264_decode_param decode_params;
+	struct mtk_h264_dpb_info h264_dpb_info[V4L2_H264_NUM_DPB_ENTRIES];
+};
+
+/**
+ * struct vdec_h264_slice_info - decode information
+ *
+ * @nal_info:		nal info of current picture
+ * @timeout:		Decode timeout: 1 timeout, 0 no timeount
+ * @bs_buf_size:	bitstream size
+ * @bs_buf_addr:	bitstream buffer dma address
+ * @y_fb_dma:		Y frame buffer dma address
+ * @c_fb_dma:		C frame buffer dma address
+ * @vdec_fb_va:	VDEC frame buffer struct virtual address
+ * @crc:		Used to check whether hardware's status is right
+ */
+struct vdec_h264_slice_info {
+	u16 nal_info;
+	u16 timeout;
+	u32 bs_buf_size;
+	u64 bs_buf_addr;
+	u64 y_fb_dma;
+	u64 c_fb_dma;
+	u64 vdec_fb_va;
+	u32 crc[8];
+};
+
+/**
+ * struct vdec_h264_slice_vsi - shared memory for decode information exchange
+ *        between SCP and Host.
+ *
+ * @wdma_err_addr:		wdma error dma address
+ * @wdma_start_addr:		wdma start dma address
+ * @wdma_end_addr:		wdma end dma address
+ * @slice_bc_start_addr:	slice bc start dma address
+ * @slice_bc_end_addr:		slice bc end dma address
+ * @row_info_start_addr:	row info start dma address
+ * @row_info_end_addr:		row info end dma address
+ * @trans_start:		trans start dma address
+ * @trans_end:			trans end dma address
+ * @wdma_end_addr_offset:	wdma end address offset
+ *
+ * @mv_buf_dma:		HW working motion vector buffer
+ *				dma address (AP-W, VPU-R)
+ * @dec:			decode information (AP-R, VPU-W)
+ * @h264_slice_params:		decode parameters for hw used
+ */
+struct vdec_h264_slice_vsi {
+	/* LAT dec addr */
+	u64 wdma_err_addr;
+	u64 wdma_start_addr;
+	u64 wdma_end_addr;
+	u64 slice_bc_start_addr;
+	u64 slice_bc_end_addr;
+	u64 row_info_start_addr;
+	u64 row_info_end_addr;
+	u64 trans_start;
+	u64 trans_end;
+	u64 wdma_end_addr_offset;
+
+	u64 mv_buf_dma[H264_MAX_MV_NUM];
+	struct vdec_h264_slice_info dec;
+	struct vdec_h264_slice_lat_dec_param h264_slice_params;
+};
+
+/**
+ * struct vdec_h264_slice_share_info - shared information used to exchange
+ *                                     message between lat and core
+ *
+ * @sps:		sequence header information from user space
+ * @dec_params:	decoder params from user space
+ * @h264_slice_params:	decoder params used for hardware
+ * @trans_start:	trans start dma address
+ * @trans_end:		trans end dma address
+ * @nal_info:		nal info of current picture
+ */
+struct vdec_h264_slice_share_info {
+	struct v4l2_ctrl_h264_sps sps;
+	struct v4l2_ctrl_h264_decode_params dec_params;
+	struct vdec_h264_slice_lat_dec_param h264_slice_params;
+	u64 trans_start;
+	u64 trans_end;
+	u16 nal_info;
+};
+
+/**
+ * struct vdec_h264_slice_inst - h264 decoder instance
+ *
+ * @slice_dec_num:	how many picture be decoded
+ * @ctx:		point to mtk_vcodec_dec_ctx
+ * @pred_buf:		HW working predication buffer
+ * @mv_buf:		HW working motion vector buffer
+ * @vpu:		VPU instance
+ * @vsi:		vsi used for lat
+ * @vsi_core:		vsi used for core
+ *
+ * @vsi_ctx:		Local VSI data for this decoding context
+ * @h264_slice_param:	the parameters that hardware use to decode
+ *
+ * @resolution_changed:resolution changed
+ * @realloc_mv_buf:	reallocate mv buffer
+ * @cap_num_planes:	number of capture queue plane
+ *
+ * @dpb:		decoded picture buffer used to store reference
+ *			buffer information
+ *@is_field_bitstream:	is field bitstream
+ */
+struct vdec_h264_slice_inst {
+	unsigned int slice_dec_num;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct mtk_vcodec_mem pred_buf;
+	struct mtk_vcodec_mem mv_buf[H264_MAX_MV_NUM];
+	struct vdec_vpu_inst vpu;
+	struct vdec_h264_slice_vsi *vsi;
+	struct vdec_h264_slice_vsi *vsi_core;
+
+	struct vdec_h264_slice_vsi vsi_ctx;
+	struct vdec_h264_slice_lat_dec_param h264_slice_param;
+
+	unsigned int resolution_changed;
+	unsigned int realloc_mv_buf;
+	unsigned int cap_num_planes;
+
+	struct v4l2_h264_dpb_entry dpb[16];
+	bool is_field_bitstream;
+};
+
+static int vdec_h264_slice_fill_decode_parameters(struct vdec_h264_slice_inst *inst,
+						  struct vdec_h264_slice_share_info *share_info)
+{
+	struct vdec_h264_slice_lat_dec_param *slice_param = &inst->vsi->h264_slice_params;
+	const struct v4l2_ctrl_h264_decode_params *dec_params;
+	const struct v4l2_ctrl_h264_scaling_matrix *src_matrix;
+	const struct v4l2_ctrl_h264_sps *sps;
+	const struct v4l2_ctrl_h264_pps *pps;
+
+	dec_params =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS);
+	if (IS_ERR(dec_params))
+		return PTR_ERR(dec_params);
+
+	src_matrix =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX);
+	if (IS_ERR(src_matrix))
+		return PTR_ERR(src_matrix);
+
+	sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS);
+	if (IS_ERR(sps))
+		return PTR_ERR(sps);
+
+	pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS);
+	if (IS_ERR(pps))
+		return PTR_ERR(pps);
+
+	if (dec_params->flags & V4L2_H264_DECODE_PARAM_FLAG_FIELD_PIC) {
+		mtk_vdec_err(inst->ctx, "No support for H.264 field decoding.");
+		inst->is_field_bitstream = true;
+		return -EINVAL;
+	}
+
+	mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps);
+	mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps);
+	mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, src_matrix);
+
+	memcpy(&share_info->sps, sps, sizeof(*sps));
+	memcpy(&share_info->dec_params, dec_params, sizeof(*dec_params));
+
+	return 0;
+}
+
+static int get_vdec_sig_decode_parameters(struct vdec_h264_slice_inst *inst)
+{
+	const struct v4l2_ctrl_h264_decode_params *dec_params;
+	const struct v4l2_ctrl_h264_sps *sps;
+	const struct v4l2_ctrl_h264_pps *pps;
+	const struct v4l2_ctrl_h264_scaling_matrix *scaling_matrix;
+	struct vdec_h264_slice_lat_dec_param *slice_param = &inst->h264_slice_param;
+	struct v4l2_h264_reflist_builder reflist_builder;
+	struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN];
+	u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0;
+	u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0;
+	u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1;
+
+	dec_params =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_DECODE_PARAMS);
+	if (IS_ERR(dec_params))
+		return PTR_ERR(dec_params);
+
+	sps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SPS);
+	if (IS_ERR(sps))
+		return PTR_ERR(sps);
+
+	pps = mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_PPS);
+	if (IS_ERR(pps))
+		return PTR_ERR(pps);
+
+	scaling_matrix =
+		mtk_vdec_h264_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_H264_SCALING_MATRIX);
+	if (IS_ERR(scaling_matrix))
+		return PTR_ERR(scaling_matrix);
+
+	mtk_vdec_h264_update_dpb(dec_params, inst->dpb);
+
+	mtk_vdec_h264_copy_sps_params(&slice_param->sps, sps);
+	mtk_vdec_h264_copy_pps_params(&slice_param->pps, pps);
+	mtk_vdec_h264_copy_scaling_matrix(&slice_param->scaling_matrix, scaling_matrix);
+
+	mtk_vdec_h264_copy_decode_params(&slice_param->decode_params, dec_params, inst->dpb);
+	mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params,
+				    slice_param->h264_dpb_info);
+
+	/* Build the reference lists */
+	v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps, inst->dpb);
+	v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist);
+	v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist, v4l2_b1_reflist);
+
+	/* Adapt the built lists to the firmware's expectations */
+	mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid);
+
+	memcpy(&inst->vsi_ctx.h264_slice_params, slice_param,
+	       sizeof(inst->vsi_ctx.h264_slice_params));
+
+	return 0;
+}
+
+static void vdec_h264_slice_fill_decode_reflist(struct vdec_h264_slice_inst *inst,
+						struct vdec_h264_slice_lat_dec_param *slice_param,
+						struct vdec_h264_slice_share_info *share_info)
+{
+	struct v4l2_ctrl_h264_decode_params *dec_params = &share_info->dec_params;
+	struct v4l2_ctrl_h264_sps *sps = &share_info->sps;
+	struct v4l2_h264_reflist_builder reflist_builder;
+	struct v4l2_h264_reference v4l2_p0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b0_reflist[V4L2_H264_REF_LIST_LEN];
+	struct v4l2_h264_reference v4l2_b1_reflist[V4L2_H264_REF_LIST_LEN];
+	u8 *p0_reflist = slice_param->decode_params.ref_pic_list_p0;
+	u8 *b0_reflist = slice_param->decode_params.ref_pic_list_b0;
+	u8 *b1_reflist = slice_param->decode_params.ref_pic_list_b1;
+
+	mtk_vdec_h264_update_dpb(dec_params, inst->dpb);
+
+	mtk_vdec_h264_copy_decode_params(&slice_param->decode_params, dec_params,
+					 inst->dpb);
+	mtk_vdec_h264_fill_dpb_info(inst->ctx, &slice_param->decode_params,
+				    slice_param->h264_dpb_info);
+
+	mtk_v4l2_vdec_dbg(3, inst->ctx, "cur poc = %d\n", dec_params->bottom_field_order_cnt);
+	/* Build the reference lists */
+	v4l2_h264_init_reflist_builder(&reflist_builder, dec_params, sps,
+				       inst->dpb);
+	v4l2_h264_build_p_ref_list(&reflist_builder, v4l2_p0_reflist);
+	v4l2_h264_build_b_ref_lists(&reflist_builder, v4l2_b0_reflist, v4l2_b1_reflist);
+
+	/* Adapt the built lists to the firmware's expectations */
+	mtk_vdec_h264_get_ref_list(p0_reflist, v4l2_p0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b0_reflist, v4l2_b0_reflist, reflist_builder.num_valid);
+	mtk_vdec_h264_get_ref_list(b1_reflist, v4l2_b1_reflist, reflist_builder.num_valid);
+}
+
+static int vdec_h264_slice_alloc_mv_buf(struct vdec_h264_slice_inst *inst,
+					struct vdec_pic_info *pic)
+{
+	unsigned int buf_sz = mtk_vdec_h264_get_mv_buf_size(pic->buf_w, pic->buf_h);
+	struct mtk_vcodec_mem *mem;
+	int i, err;
+
+	mtk_v4l2_vdec_dbg(3, inst->ctx, "size = 0x%x", buf_sz);
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+		mem->size = buf_sz;
+		err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+		if (err) {
+			mtk_vdec_err(inst->ctx, "failed to allocate mv buf");
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static void vdec_h264_slice_free_mv_buf(struct vdec_h264_slice_inst *inst)
+{
+	int i;
+	struct mtk_vcodec_mem *mem;
+
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+	}
+}
+
+static void vdec_h264_slice_get_pic_info(struct vdec_h264_slice_inst *inst)
+{
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+	u32 data[3];
+
+	data[0] = ctx->picinfo.pic_w;
+	data[1] = ctx->picinfo.pic_h;
+	data[2] = ctx->capture_fourcc;
+	vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO);
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0];
+	ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1];
+	inst->cap_num_planes =
+		ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes;
+
+	mtk_vdec_debug(ctx, "pic(%d, %d), buf(%d, %d)",
+		       ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+		       ctx->picinfo.buf_w, ctx->picinfo.buf_h);
+	mtk_vdec_debug(ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0],
+		       ctx->picinfo.fb_sz[1]);
+
+	if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w ||
+	    ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) {
+		inst->resolution_changed = true;
+		if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w ||
+		    ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h)
+			inst->realloc_mv_buf = true;
+
+		mtk_v4l2_vdec_dbg(1, inst->ctx, "resChg: (%d %d) : old(%d, %d) -> new(%d, %d)",
+				  inst->resolution_changed,
+				  inst->realloc_mv_buf,
+				  ctx->last_decoded_picinfo.pic_w,
+				  ctx->last_decoded_picinfo.pic_h,
+				  ctx->picinfo.pic_w, ctx->picinfo.pic_h);
+	}
+}
+
+static void vdec_h264_slice_get_crop_info(struct vdec_h264_slice_inst *inst,
+					  struct v4l2_rect *cr)
+{
+	cr->left = 0;
+	cr->top = 0;
+	cr->width = inst->ctx->picinfo.pic_w;
+	cr->height = inst->ctx->picinfo.pic_h;
+
+	mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static int vdec_h264_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_h264_slice_inst *inst;
+	int err, vsi_size;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = SCP_IPI_VDEC_LAT;
+	inst->vpu.core_id = SCP_IPI_VDEC_CORE;
+	inst->vpu.ctx = ctx;
+	inst->vpu.codec_type = ctx->current_codec;
+	inst->vpu.capture_type = ctx->capture_fourcc;
+
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_h264 init err=%d", err);
+		goto error_free_inst;
+	}
+
+	vsi_size = round_up(sizeof(struct vdec_h264_slice_vsi), VCODEC_DEC_ALIGNED_64);
+	inst->vsi = inst->vpu.vsi;
+	inst->vsi_core =
+		(struct vdec_h264_slice_vsi *)(((char *)inst->vpu.vsi) + vsi_size);
+	inst->resolution_changed = true;
+	inst->realloc_mv_buf = true;
+
+	mtk_vdec_debug(ctx, "lat struct size = %d,%d,%d,%d vsi: %d\n",
+		       (int)sizeof(struct mtk_h264_sps_param),
+		       (int)sizeof(struct mtk_h264_pps_param),
+		       (int)sizeof(struct vdec_h264_slice_lat_dec_param),
+		       (int)sizeof(struct mtk_h264_dpb_info),
+		       vsi_size);
+	mtk_vdec_debug(ctx, "lat H264 instance >> %p, codec_type = 0x%x",
+		       inst, inst->vpu.codec_type);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static void vdec_h264_slice_deinit(void *h_vdec)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+
+	vpu_dec_deinit(&inst->vpu);
+	vdec_h264_slice_free_mv_buf(inst);
+	vdec_msg_queue_deinit(&inst->ctx->msg_queue, inst->ctx);
+
+	kfree(inst);
+}
+
+static int vdec_h264_slice_core_decode(struct vdec_lat_buf *lat_buf)
+{
+	struct vdec_fb *fb;
+	u64 vdec_fb_va;
+	u64 y_fb_dma, c_fb_dma;
+	int err, timeout, i;
+	struct mtk_vcodec_dec_ctx *ctx = lat_buf->ctx;
+	struct vdec_h264_slice_inst *inst = ctx->drv_handle;
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	struct vdec_h264_slice_share_info *share_info = lat_buf->private_data;
+	struct mtk_vcodec_mem *mem;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+
+	mtk_vdec_debug(ctx, "[h264-core] vdec_h264 core decode");
+	memcpy(&inst->vsi_core->h264_slice_params, &share_info->h264_slice_params,
+	       sizeof(share_info->h264_slice_params));
+
+	fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx);
+	if (!fb) {
+		err = -EBUSY;
+		mtk_vdec_err(ctx, "fb buffer is NULL");
+		goto vdec_dec_end;
+	}
+
+	vdec_fb_va = (unsigned long)fb;
+	y_fb_dma = (u64)fb->base_y.dma_addr;
+	if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1)
+		c_fb_dma =
+			y_fb_dma + inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h;
+	else
+		c_fb_dma = (u64)fb->base_c.dma_addr;
+
+	mtk_vdec_debug(ctx, "[h264-core] y/c addr = 0x%llx 0x%llx", y_fb_dma, c_fb_dma);
+
+	inst->vsi_core->dec.y_fb_dma = y_fb_dma;
+	inst->vsi_core->dec.c_fb_dma = c_fb_dma;
+	inst->vsi_core->dec.vdec_fb_va = vdec_fb_va;
+	inst->vsi_core->dec.nal_info = share_info->nal_info;
+	inst->vsi_core->wdma_start_addr =
+		lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	inst->vsi_core->wdma_end_addr =
+		lat_buf->ctx->msg_queue.wdma_addr.dma_addr +
+		lat_buf->ctx->msg_queue.wdma_addr.size;
+	inst->vsi_core->wdma_err_addr = lat_buf->wdma_err_addr.dma_addr;
+	inst->vsi_core->slice_bc_start_addr = lat_buf->slice_bc_addr.dma_addr;
+	inst->vsi_core->slice_bc_end_addr = lat_buf->slice_bc_addr.dma_addr +
+		lat_buf->slice_bc_addr.size;
+	inst->vsi_core->trans_start = share_info->trans_start;
+	inst->vsi_core->trans_end = share_info->trans_end;
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		inst->vsi_core->mv_buf_dma[i] = mem->dma_addr;
+	}
+
+	vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx);
+	v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, vb2_v4l2, true);
+
+	vdec_h264_slice_fill_decode_reflist(inst, &inst->vsi_core->h264_slice_params,
+					    share_info);
+
+	err = vpu_dec_core(vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "core decode err=%d", err);
+		goto vdec_dec_end;
+	}
+
+	/* wait decoder done interrupt */
+	timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					       WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE);
+	if (timeout)
+		mtk_vdec_err(ctx, "core decode timeout: pic_%d", ctx->decoded_frame_cnt);
+	inst->vsi_core->dec.timeout = !!timeout;
+
+	vpu_dec_core_end(vpu);
+	mtk_vdec_debug(ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x",
+		       ctx->decoded_frame_cnt,
+		       inst->vsi_core->dec.crc[0], inst->vsi_core->dec.crc[1],
+		       inst->vsi_core->dec.crc[2], inst->vsi_core->dec.crc[3],
+		       inst->vsi_core->dec.crc[4], inst->vsi_core->dec.crc[5],
+		       inst->vsi_core->dec.crc[6], inst->vsi_core->dec.crc[7]);
+
+vdec_dec_end:
+	vdec_msg_queue_update_ube_rptr(&lat_buf->ctx->msg_queue, share_info->trans_end);
+	ctx->dev->vdec_pdata->cap_to_disp(ctx, !!err, lat_buf->src_buf_req);
+	mtk_vdec_debug(ctx, "core decode done err=%d", err);
+	ctx->decoded_frame_cnt++;
+	return 0;
+}
+
+static void vdec_h264_insert_startcode(struct mtk_vcodec_dec_dev *vcodec_dev, unsigned char *buf,
+				       size_t *bs_size, struct mtk_h264_pps_param *pps)
+{
+	struct device *dev = &vcodec_dev->plat_dev->dev;
+
+	/* Need to add pending data at the end of bitstream when bs_sz is small than
+	 * 20 bytes for cavlc bitstream, or lat will decode fail. This pending data is
+	 * useful for mt8192 and mt8195 platform.
+	 *
+	 * cavlc bitstream when entropy_coding_mode_flag is false.
+	 */
+	if (pps->entropy_coding_mode_flag || *bs_size > 20 ||
+	    !(of_device_is_compatible(dev->of_node, "mediatek,mt8192-vcodec-dec") ||
+	    of_device_is_compatible(dev->of_node, "mediatek,mt8195-vcodec-dec")))
+		return;
+
+	buf[*bs_size] = 0;
+	buf[*bs_size + 1] = 0;
+	buf[*bs_size + 2] = 1;
+	buf[*bs_size + 3] = 0xff;
+	(*bs_size) += 4;
+}
+
+static int vdec_h264_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				      struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	struct mtk_video_dec_buf *src_buf_info;
+	int nal_start_idx, err, timeout = 0, i;
+	unsigned int data[2];
+	struct vdec_lat_buf *lat_buf;
+	struct vdec_h264_slice_share_info *share_info;
+	unsigned char *buf;
+	struct mtk_vcodec_mem *mem;
+
+	if (vdec_msg_queue_init(&inst->ctx->msg_queue, inst->ctx,
+				vdec_h264_slice_core_decode,
+				sizeof(*share_info)))
+		return -ENOMEM;
+
+	/* bs NULL means flush decoder */
+	if (!bs) {
+		vdec_msg_queue_wait_lat_buf_full(&inst->ctx->msg_queue);
+		return vpu_dec_reset(vpu);
+	}
+
+	if (inst->is_field_bitstream)
+		return -EINVAL;
+
+	lat_buf = vdec_msg_queue_dqbuf(&inst->ctx->msg_queue.lat_ctx);
+	if (!lat_buf) {
+		mtk_vdec_debug(inst->ctx, "failed to get lat buffer");
+		return -EAGAIN;
+	}
+	share_info = lat_buf->private_data;
+	src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer);
+
+	buf = (unsigned char *)bs->va;
+	nal_start_idx = mtk_vdec_h264_find_start_code(buf, bs->size);
+	if (nal_start_idx < 0) {
+		err = -EINVAL;
+		goto err_free_fb_out;
+	}
+
+	inst->vsi->dec.nal_info = buf[nal_start_idx];
+	lat_buf->src_buf_req = src_buf_info->m2m_buf.vb.vb2_buf.req_obj.req;
+	v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, &lat_buf->ts_info, true);
+
+	err = vdec_h264_slice_fill_decode_parameters(inst, share_info);
+	if (err)
+		goto err_free_fb_out;
+
+	vdec_h264_insert_startcode(inst->ctx->dev, buf, &bs->size,
+				   &share_info->h264_slice_params.pps);
+
+	inst->vsi->dec.bs_buf_addr = (uint64_t)bs->dma_addr;
+	inst->vsi->dec.bs_buf_size = bs->size;
+
+	*res_chg = inst->resolution_changed;
+	if (inst->resolution_changed) {
+		mtk_vdec_debug(inst->ctx, "- resolution changed -");
+		if (inst->realloc_mv_buf) {
+			err = vdec_h264_slice_alloc_mv_buf(inst, &inst->ctx->picinfo);
+			inst->realloc_mv_buf = false;
+			if (err)
+				goto err_free_fb_out;
+		}
+		inst->resolution_changed = false;
+	}
+	for (i = 0; i < H264_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		inst->vsi->mv_buf_dma[i] = mem->dma_addr;
+	}
+	inst->vsi->wdma_start_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	inst->vsi->wdma_end_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr +
+		lat_buf->ctx->msg_queue.wdma_addr.size;
+	inst->vsi->wdma_err_addr = lat_buf->wdma_err_addr.dma_addr;
+	inst->vsi->slice_bc_start_addr = lat_buf->slice_bc_addr.dma_addr;
+	inst->vsi->slice_bc_end_addr = lat_buf->slice_bc_addr.dma_addr +
+		lat_buf->slice_bc_addr.size;
+
+	inst->vsi->trans_end = inst->ctx->msg_queue.wdma_rptr_addr;
+	inst->vsi->trans_start = inst->ctx->msg_queue.wdma_wptr_addr;
+	mtk_vdec_debug(inst->ctx, "lat:trans(0x%llx 0x%llx) err:0x%llx",
+		       inst->vsi->wdma_start_addr,
+		       inst->vsi->wdma_end_addr,
+		       inst->vsi->wdma_err_addr);
+
+	mtk_vdec_debug(inst->ctx, "slice(0x%llx 0x%llx) rprt((0x%llx 0x%llx))",
+		       inst->vsi->slice_bc_start_addr,
+		       inst->vsi->slice_bc_end_addr,
+		       inst->vsi->trans_start,
+		       inst->vsi->trans_end);
+	err = vpu_dec_start(vpu, data, 2);
+	if (err) {
+		mtk_vdec_debug(inst->ctx, "lat decode err: %d", err);
+		goto err_free_fb_out;
+	}
+
+	share_info->trans_end = inst->ctx->msg_queue.wdma_addr.dma_addr +
+		inst->vsi->wdma_end_addr_offset;
+	share_info->trans_start = inst->ctx->msg_queue.wdma_wptr_addr;
+	share_info->nal_info = inst->vsi->dec.nal_info;
+
+	if (IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) {
+		memcpy(&share_info->h264_slice_params, &inst->vsi->h264_slice_params,
+		       sizeof(share_info->h264_slice_params));
+		vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf);
+	}
+
+	/* wait decoder done interrupt */
+	timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					       WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0);
+	if (timeout)
+		mtk_vdec_err(inst->ctx, "lat decode timeout: pic_%d", inst->slice_dec_num);
+	inst->vsi->dec.timeout = !!timeout;
+
+	err = vpu_dec_end(vpu);
+	if (err == SLICE_HEADER_FULL || err == TRANS_BUFFER_FULL) {
+		if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability))
+			vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf);
+		inst->slice_dec_num++;
+		mtk_vdec_err(inst->ctx, "lat dec fail: pic_%d err:%d", inst->slice_dec_num, err);
+		return -EINVAL;
+	}
+
+	share_info->trans_end = inst->ctx->msg_queue.wdma_addr.dma_addr +
+		inst->vsi->wdma_end_addr_offset;
+	vdec_msg_queue_update_ube_wptr(&lat_buf->ctx->msg_queue, share_info->trans_end);
+
+	if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) {
+		memcpy(&share_info->h264_slice_params, &inst->vsi->h264_slice_params,
+		       sizeof(share_info->h264_slice_params));
+		vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf);
+	}
+	mtk_vdec_debug(inst->ctx, "dec num: %d lat crc: 0x%x 0x%x 0x%x", inst->slice_dec_num,
+		       inst->vsi->dec.crc[0], inst->vsi->dec.crc[1], inst->vsi->dec.crc[2]);
+
+	inst->slice_dec_num++;
+	return 0;
+err_free_fb_out:
+	vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf);
+	mtk_vdec_err(inst->ctx, "slice dec number: %d err: %d", inst->slice_dec_num, err);
+	return err;
+}
+
+static int vdec_h264_slice_single_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+					 struct vdec_fb *unused, bool *res_chg)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	struct mtk_video_dec_buf *src_buf_info, *dst_buf_info;
+	struct vdec_fb *fb;
+	unsigned char *buf;
+	unsigned int data[2], i;
+	u64 y_fb_dma, c_fb_dma;
+	struct mtk_vcodec_mem *mem;
+	int err, nal_start_idx;
+
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vpu_dec_reset(vpu);
+
+	fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx);
+	src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer);
+	dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer);
+
+	y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0;
+	c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0;
+	mtk_vdec_debug(inst->ctx, "[h264-dec] [%d] y_dma=%llx c_dma=%llx",
+		       inst->ctx->decoded_frame_cnt, y_fb_dma, c_fb_dma);
+
+	inst->vsi_ctx.dec.bs_buf_addr = (u64)bs->dma_addr;
+	inst->vsi_ctx.dec.bs_buf_size = bs->size;
+	inst->vsi_ctx.dec.y_fb_dma = y_fb_dma;
+	inst->vsi_ctx.dec.c_fb_dma = c_fb_dma;
+	inst->vsi_ctx.dec.vdec_fb_va = (u64)(uintptr_t)fb;
+
+	v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb,
+				   &dst_buf_info->m2m_buf.vb, true);
+	err = get_vdec_sig_decode_parameters(inst);
+	if (err)
+		goto err_free_fb_out;
+
+	buf = (unsigned char *)bs->va;
+	nal_start_idx = mtk_vdec_h264_find_start_code(buf, bs->size);
+	if (nal_start_idx < 0) {
+		err = -EINVAL;
+		goto err_free_fb_out;
+	}
+	inst->vsi_ctx.dec.nal_info = buf[nal_start_idx];
+
+	*res_chg = inst->resolution_changed;
+	if (inst->resolution_changed) {
+		mtk_vdec_debug(inst->ctx, "- resolution changed -");
+		if (inst->realloc_mv_buf) {
+			err = vdec_h264_slice_alloc_mv_buf(inst, &inst->ctx->picinfo);
+			inst->realloc_mv_buf = false;
+			if (err)
+				goto err_free_fb_out;
+		}
+		inst->resolution_changed = false;
+
+		for (i = 0; i < H264_MAX_MV_NUM; i++) {
+			mem = &inst->mv_buf[i];
+			inst->vsi_ctx.mv_buf_dma[i] = mem->dma_addr;
+		}
+	}
+
+	memcpy(inst->vpu.vsi, &inst->vsi_ctx, sizeof(inst->vsi_ctx));
+	err = vpu_dec_start(vpu, data, 2);
+	if (err)
+		goto err_free_fb_out;
+
+	/* wait decoder done interrupt */
+	err = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					   WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE);
+	if (err)
+		mtk_vdec_err(inst->ctx, "decode timeout: pic_%d", inst->ctx->decoded_frame_cnt);
+
+	inst->vsi->dec.timeout = !!err;
+	err = vpu_dec_end(vpu);
+	if (err)
+		goto err_free_fb_out;
+
+	memcpy(&inst->vsi_ctx, inst->vpu.vsi, sizeof(inst->vsi_ctx));
+	mtk_vdec_debug(inst->ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x",
+		       inst->ctx->decoded_frame_cnt,
+		       inst->vsi_ctx.dec.crc[0], inst->vsi_ctx.dec.crc[1],
+		       inst->vsi_ctx.dec.crc[2], inst->vsi_ctx.dec.crc[3],
+		       inst->vsi_ctx.dec.crc[4], inst->vsi_ctx.dec.crc[5],
+		       inst->vsi_ctx.dec.crc[6], inst->vsi_ctx.dec.crc[7]);
+
+	inst->ctx->decoded_frame_cnt++;
+	return 0;
+
+err_free_fb_out:
+	mtk_vdec_err(inst->ctx, "dec frame number: %d err: %d", inst->ctx->decoded_frame_cnt, err);
+	return err;
+}
+
+static int vdec_h264_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				  struct vdec_fb *unused, bool *res_chg)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+	int ret;
+
+	if (!h_vdec)
+		return -EINVAL;
+
+	if (inst->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE)
+		ret = vdec_h264_slice_single_decode(h_vdec, bs, unused, res_chg);
+	else
+		ret = vdec_h264_slice_lat_decode(h_vdec, bs, unused, res_chg);
+
+	return ret;
+}
+
+static int vdec_h264_slice_get_param(void *h_vdec, enum vdec_get_param_type type,
+				     void *out)
+{
+	struct vdec_h264_slice_inst *inst = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		vdec_h264_slice_get_pic_info(inst);
+		break;
+	case GET_PARAM_DPB_SIZE:
+		*(unsigned int *)out = 6;
+		break;
+	case GET_PARAM_CROP_INFO:
+		vdec_h264_slice_get_crop_info(inst, out);
+		break;
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+const struct vdec_common_if vdec_h264_slice_multi_if = {
+	.init		= vdec_h264_slice_init,
+	.decode		= vdec_h264_slice_decode,
+	.get_param	= vdec_h264_slice_get_param,
+	.deinit		= vdec_h264_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c
new file mode 100644
index 0000000000..06ed47df69
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c
@@ -0,0 +1,1092 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2023 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/videobuf2-dma-contig.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+
+/* the size used to store hevc wrap information */
+#define VDEC_HEVC_WRAP_SZ (532 * SZ_1K)
+
+#define HEVC_MAX_MV_NUM 32
+
+/* get used parameters for sps/pps */
+#define GET_HEVC_VDEC_FLAG(cond, flag) \
+	{ dst_param->cond = ((src_param->flags & (flag)) ? (1) : (0)); }
+#define GET_HEVC_VDEC_PARAM(param) \
+	{ dst_param->param = src_param->param; }
+
+/**
+ * enum vdec_hevc_core_dec_err_type  - core decode error type
+ *
+ * @TRANS_BUFFER_FULL: trans buffer is full
+ * @SLICE_HEADER_FULL: slice header buffer is full
+ */
+enum vdec_hevc_core_dec_err_type {
+	TRANS_BUFFER_FULL = 1,
+	SLICE_HEADER_FULL,
+};
+
+/**
+ * struct mtk_hevc_dpb_info  - hevc dpb information
+ *
+ * @y_dma_addr:     Y plane physical address
+ * @c_dma_addr:     CbCr plane physical address
+ * @reference_flag: reference picture flag (short/long term reference picture)
+ * @field:          field picture flag
+ */
+struct mtk_hevc_dpb_info {
+	dma_addr_t y_dma_addr;
+	dma_addr_t c_dma_addr;
+	int reference_flag;
+	int field;
+};
+
+/*
+ * struct mtk_hevc_sps_param  - parameters for sps
+ */
+struct mtk_hevc_sps_param {
+	unsigned char video_parameter_set_id;
+	unsigned char seq_parameter_set_id;
+	unsigned short pic_width_in_luma_samples;
+	unsigned short pic_height_in_luma_samples;
+	unsigned char bit_depth_luma_minus8;
+	unsigned char bit_depth_chroma_minus8;
+	unsigned char log2_max_pic_order_cnt_lsb_minus4;
+	unsigned char sps_max_dec_pic_buffering_minus1;
+	unsigned char sps_max_num_reorder_pics;
+	unsigned char sps_max_latency_increase_plus1;
+	unsigned char log2_min_luma_coding_block_size_minus3;
+	unsigned char log2_diff_max_min_luma_coding_block_size;
+	unsigned char log2_min_luma_transform_block_size_minus2;
+	unsigned char log2_diff_max_min_luma_transform_block_size;
+	unsigned char max_transform_hierarchy_depth_inter;
+	unsigned char max_transform_hierarchy_depth_intra;
+	unsigned char pcm_sample_bit_depth_luma_minus1;
+	unsigned char pcm_sample_bit_depth_chroma_minus1;
+	unsigned char log2_min_pcm_luma_coding_block_size_minus3;
+	unsigned char log2_diff_max_min_pcm_luma_coding_block_size;
+	unsigned char num_short_term_ref_pic_sets;
+	unsigned char num_long_term_ref_pics_sps;
+	unsigned char chroma_format_idc;
+	unsigned char sps_max_sub_layers_minus1;
+	unsigned char separate_colour_plane;
+	unsigned char scaling_list_enabled;
+	unsigned char amp_enabled;
+	unsigned char sample_adaptive_offset;
+	unsigned char pcm_enabled;
+	unsigned char pcm_loop_filter_disabled;
+	unsigned char long_term_ref_pics_enabled;
+	unsigned char sps_temporal_mvp_enabled;
+	unsigned char strong_intra_smoothing_enabled;
+	unsigned char reserved[5];
+};
+
+/*
+ * struct mtk_hevc_pps_param  - parameters for pps
+ */
+struct mtk_hevc_pps_param {
+	unsigned char pic_parameter_set_id;
+	unsigned char num_extra_slice_header_bits;
+	unsigned char num_ref_idx_l0_default_active_minus1;
+	unsigned char num_ref_idx_l1_default_active_minus1;
+	char init_qp_minus26;
+	unsigned char diff_cu_qp_delta_depth;
+	char pps_cb_qp_offset;
+	char pps_cr_qp_offset;
+	unsigned char num_tile_columns_minus1;
+	unsigned char num_tile_rows_minus1;
+	unsigned char column_width_minus1[20];
+	unsigned char row_height_minus1[22];
+	char pps_beta_offset_div2;
+	char pps_tc_offset_div2;
+	unsigned char log2_parallel_merge_level_minus2;
+	char dependent_slice_segment_enabled;
+	char output_flag_present;
+	char sign_data_hiding_enabled;
+	char cabac_init_present;
+	char constrained_intra_pred;
+	char transform_skip_enabled;
+	char cu_qp_delta_enabled;
+	char pps_slice_chroma_qp_offsets_present;
+	char weighted_pred;
+	char weighted_bipred;
+	char transquant_bypass_enabled;
+	char pps_flag_tiles_enabled;
+	char entropy_coding_sync_enabled;
+	char loop_filter_across_tiles_enabled;
+	char pps_loop_filter_across_slices_enabled;
+	char deblocking_filter_override_enabled;
+	char pps_disable_deflocking_filter;
+	char lists_modification_present;
+	char slice_segment_header_extersion_present;
+	char deblocking_filter_control_present;
+	char uniform_spacing;
+	char reserved[6];
+};
+
+/*
+ * struct mtk_hevc_slice_header_param  - parameters for slice header
+ */
+struct mtk_hevc_slice_header_param {
+	unsigned int	slice_type;
+	unsigned int	num_active_ref_layer_pics;
+	int		slice_qp;
+	int		slice_qp_delta_cb;
+	int		slice_qp_delta_cr;
+	int		num_ref_idx[3];
+	unsigned int	col_ref_idx;
+	unsigned int	five_minus_max_num_merge_cand;
+	int		slice_deblocking_filter_beta_offset_div2;
+	int		slice_deblocking_filter_tc_offset_div2;
+	unsigned char	sao_enable_flag;
+	unsigned char	sao_enable_flag_chroma;
+	unsigned char	cabac_init_flag;
+	unsigned char	slice_tmvp_flags_present;
+	unsigned char	col_from_l0_flag;
+	unsigned char	mvd_l1_zero_flag;
+	unsigned char	slice_loop_filter_across_slices_enabled_flag;
+	unsigned char	deblocking_filter_disable_flag;
+	unsigned int	slice_reg0;
+	unsigned int	slice_reg1;
+	unsigned int	slice_reg2;
+	unsigned int	num_rps_curr_temp_list;
+	unsigned int	ref_list_mode;
+	int		str_num_delta_pocs;
+	int		str_num_negtive_pos_pics;
+	int		num_long_term;
+	int		num_long_term_sps;
+	unsigned int	max_cu_width;
+	unsigned int	max_cu_height;
+	unsigned int	num_entry_point_offsets;
+	unsigned int    last_lcu_x_in_tile[17];
+	unsigned int    last_lcu_y_in_tile[17];
+	unsigned char   nal_unit_type;
+};
+
+/*
+ * struct slice_api_hevc_scaling_matrix  - parameters for scaling list
+ */
+struct slice_api_hevc_scaling_matrix {
+	unsigned char scaling_list_4x4[6][16];
+	unsigned char scaling_list_8x8[6][64];
+	unsigned char scaling_list_16x16[6][64];
+	unsigned char scaling_list_32x32[2][64];
+	unsigned char scaling_list_dc_coef_16x16[6];
+	unsigned char scaling_list_dc_coef_32x32[2];
+};
+
+/*
+ * struct slice_hevc_dpb_entry  - each dpb information
+ */
+struct slice_hevc_dpb_entry {
+	u64 timestamp;
+	unsigned char flags;
+	unsigned char field_pic;
+	int pic_order_cnt_val;
+};
+
+/*
+ * struct slice_api_hevc_decode_param - parameters for decode.
+ */
+struct slice_api_hevc_decode_param {
+	struct slice_hevc_dpb_entry dpb[V4L2_HEVC_DPB_ENTRIES_NUM_MAX];
+	int pic_order_cnt_val;
+	unsigned short short_term_ref_pic_set_size;
+	unsigned short long_term_ref_pic_set_size;
+	unsigned char num_active_dpb_entries;
+	unsigned char num_poc_st_curr_before;
+	unsigned char num_poc_st_curr_after;
+	unsigned char num_poc_lt_curr;
+	unsigned char poc_st_curr_before[V4L2_HEVC_DPB_ENTRIES_NUM_MAX];
+	unsigned char poc_st_curr_after[V4L2_HEVC_DPB_ENTRIES_NUM_MAX];
+	unsigned char poc_lt_curr[V4L2_HEVC_DPB_ENTRIES_NUM_MAX];
+	unsigned char num_delta_pocs_of_ref_rps_idx;
+	int flags;
+};
+
+/**
+ * struct hevc_fb - hevc decode frame buffer information
+ *
+ * @vdec_fb_va: virtual address of struct vdec_fb
+ * @y_fb_dma:   dma address of Y frame buffer (luma)
+ * @c_fb_dma:   dma address of C frame buffer (chroma)
+ * @poc:        picture order count of frame buffer
+ * @reserved:   for 8 bytes alignment
+ */
+struct hevc_fb {
+	u64 vdec_fb_va;
+	u64 y_fb_dma;
+	u64 c_fb_dma;
+	s32 poc;
+	u32 reserved;
+};
+
+/**
+ * struct vdec_hevc_slice_lat_dec_param  - parameters for decode current frame
+ *
+ * @sps:            hevc sps syntax parameters
+ * @pps:            hevc pps syntax parameters
+ * @slice_header:   hevc slice header syntax parameters
+ * @scaling_matrix: hevc scaling list parameters
+ * @decode_params:  decoder parameters of each frame used for hardware decode
+ * @hevc_dpb_info:  dpb reference list
+ */
+struct vdec_hevc_slice_lat_dec_param {
+	struct mtk_hevc_sps_param sps;
+	struct mtk_hevc_pps_param pps;
+	struct mtk_hevc_slice_header_param slice_header;
+	struct slice_api_hevc_scaling_matrix scaling_matrix;
+	struct slice_api_hevc_decode_param decode_params;
+	struct mtk_hevc_dpb_info hevc_dpb_info[V4L2_HEVC_DPB_ENTRIES_NUM_MAX];
+};
+
+/**
+ * struct vdec_hevc_slice_info - decode information
+ *
+ * @wdma_end_addr_offset: wdma end address offset
+ * @timeout:              Decode timeout: 1 timeout, 0 no timeount
+ * @vdec_fb_va:           VDEC frame buffer struct virtual address
+ * @crc:                  Used to check whether hardware's status is right
+ */
+struct vdec_hevc_slice_info {
+	u64 wdma_end_addr_offset;
+	u64 timeout;
+	u64 vdec_fb_va;
+	u32 crc[8];
+};
+
+/*
+ * struct vdec_hevc_slice_mem - memory address and size
+ */
+struct vdec_hevc_slice_mem {
+	union {
+		u64 buf;
+		dma_addr_t dma_addr;
+	};
+	union {
+		size_t size;
+		dma_addr_t dma_addr_end;
+		u64 padding;
+	};
+};
+
+/**
+ * struct vdec_hevc_slice_fb - frame buffer for decoding
+ * @y:  current y buffer address info
+ * @c:  current c buffer address info
+ */
+struct vdec_hevc_slice_fb {
+	struct vdec_hevc_slice_mem y;
+	struct vdec_hevc_slice_mem c;
+};
+
+/**
+ * struct vdec_hevc_slice_vsi - shared memory for decode information exchange
+ *        between SCP and Host.
+ *
+ * @bs:                input buffer info
+ *
+ * @ube:               ube buffer
+ * @trans:             transcoded buffer
+ * @err_map:           err map buffer
+ * @slice_bc:          slice bc buffer
+ * @wrap:              temp buffer
+ *
+ * @fb:                current y/c buffer
+ * @mv_buf_dma:        HW working motion vector buffer
+ * @dec:               decode information (AP-R, VPU-W)
+ * @hevc_slice_params: decode parameters for hw used
+ */
+struct vdec_hevc_slice_vsi {
+	/* used in LAT stage */
+	struct vdec_hevc_slice_mem bs;
+
+	struct vdec_hevc_slice_mem ube;
+	struct vdec_hevc_slice_mem trans;
+	struct vdec_hevc_slice_mem err_map;
+	struct vdec_hevc_slice_mem slice_bc;
+	struct vdec_hevc_slice_mem wrap;
+
+	struct vdec_hevc_slice_fb fb;
+	struct vdec_hevc_slice_mem mv_buf_dma[HEVC_MAX_MV_NUM];
+	struct vdec_hevc_slice_info dec;
+	struct vdec_hevc_slice_lat_dec_param hevc_slice_params;
+};
+
+/**
+ * struct vdec_hevc_slice_share_info - shared information used to exchange
+ *                                     message between lat and core
+ *
+ * @sps:               sequence header information from user space
+ * @dec_params:        decoder params from user space
+ * @hevc_slice_params: decoder params used for hardware
+ * @trans:             trans buffer dma address
+ */
+struct vdec_hevc_slice_share_info {
+	struct v4l2_ctrl_hevc_sps sps;
+	struct v4l2_ctrl_hevc_decode_params dec_params;
+	struct vdec_hevc_slice_lat_dec_param hevc_slice_params;
+	struct vdec_hevc_slice_mem trans;
+};
+
+/**
+ * struct vdec_hevc_slice_inst - hevc decoder instance
+ *
+ * @slice_dec_num:      how many picture be decoded
+ * @ctx:                point to mtk_vcodec_dec_ctx
+ * @mv_buf:             HW working motion vector buffer
+ * @vpu:                VPU instance
+ * @vsi:                vsi used for lat
+ * @vsi_core:           vsi used for core
+ * @wrap_addr:          wrap address used for hevc
+ *
+ * @hevc_slice_param:   the parameters that hardware use to decode
+ *
+ * @resolution_changed: resolution changed
+ * @realloc_mv_buf:     reallocate mv buffer
+ * @cap_num_planes:     number of capture queue plane
+ */
+struct vdec_hevc_slice_inst {
+	unsigned int slice_dec_num;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct mtk_vcodec_mem mv_buf[HEVC_MAX_MV_NUM];
+	struct vdec_vpu_inst vpu;
+	struct vdec_hevc_slice_vsi *vsi;
+	struct vdec_hevc_slice_vsi *vsi_core;
+	struct mtk_vcodec_mem wrap_addr;
+
+	struct vdec_hevc_slice_lat_dec_param hevc_slice_param;
+
+	unsigned int resolution_changed;
+	unsigned int realloc_mv_buf;
+	unsigned int cap_num_planes;
+};
+
+static unsigned int vdec_hevc_get_mv_buf_size(unsigned int width, unsigned int height)
+{
+	const unsigned int unit_size = (width / 16) * (height / 16) + 8;
+
+	return 64 * unit_size;
+}
+
+static void *vdec_hevc_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id)
+{
+	struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id);
+
+	if (!ctrl)
+		return ERR_PTR(-EINVAL);
+
+	return ctrl->p_cur.p;
+}
+
+static void vdec_hevc_fill_dpb_info(struct mtk_vcodec_dec_ctx *ctx,
+				    struct slice_api_hevc_decode_param *decode_params,
+				    struct mtk_hevc_dpb_info *hevc_dpb_info)
+{
+	const struct slice_hevc_dpb_entry *dpb;
+	struct vb2_queue *vq;
+	struct vb2_buffer *vb;
+	int index;
+
+	vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	for (index = 0; index < V4L2_HEVC_DPB_ENTRIES_NUM_MAX; index++) {
+		dpb = &decode_params->dpb[index];
+		if (index >= decode_params->num_active_dpb_entries)
+			continue;
+
+		vb = vb2_find_buffer(vq, dpb->timestamp);
+		if (!vb) {
+			dev_err(&ctx->dev->plat_dev->dev,
+				"Reference invalid: dpb_index(%d) timestamp(%lld)",
+				index, dpb->timestamp);
+			continue;
+		}
+
+		hevc_dpb_info[index].field = dpb->field_pic;
+
+		hevc_dpb_info[index].y_dma_addr = vb2_dma_contig_plane_dma_addr(vb, 0);
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2)
+			hevc_dpb_info[index].c_dma_addr = vb2_dma_contig_plane_dma_addr(vb, 1);
+		else
+			hevc_dpb_info[index].c_dma_addr =
+				hevc_dpb_info[index].y_dma_addr + ctx->picinfo.fb_sz[0];
+	}
+}
+
+static void vdec_hevc_copy_sps_params(struct mtk_hevc_sps_param *dst_param,
+				      const struct v4l2_ctrl_hevc_sps *src_param)
+{
+	GET_HEVC_VDEC_PARAM(video_parameter_set_id);
+	GET_HEVC_VDEC_PARAM(seq_parameter_set_id);
+	GET_HEVC_VDEC_PARAM(pic_width_in_luma_samples);
+	GET_HEVC_VDEC_PARAM(pic_height_in_luma_samples);
+	GET_HEVC_VDEC_PARAM(bit_depth_luma_minus8);
+	GET_HEVC_VDEC_PARAM(bit_depth_chroma_minus8);
+	GET_HEVC_VDEC_PARAM(log2_max_pic_order_cnt_lsb_minus4);
+	GET_HEVC_VDEC_PARAM(sps_max_dec_pic_buffering_minus1);
+	GET_HEVC_VDEC_PARAM(sps_max_num_reorder_pics);
+	GET_HEVC_VDEC_PARAM(sps_max_latency_increase_plus1);
+	GET_HEVC_VDEC_PARAM(log2_min_luma_coding_block_size_minus3);
+	GET_HEVC_VDEC_PARAM(log2_diff_max_min_luma_coding_block_size);
+	GET_HEVC_VDEC_PARAM(log2_min_luma_transform_block_size_minus2);
+	GET_HEVC_VDEC_PARAM(log2_diff_max_min_luma_transform_block_size);
+	GET_HEVC_VDEC_PARAM(max_transform_hierarchy_depth_inter);
+	GET_HEVC_VDEC_PARAM(max_transform_hierarchy_depth_intra);
+	GET_HEVC_VDEC_PARAM(pcm_sample_bit_depth_luma_minus1);
+	GET_HEVC_VDEC_PARAM(pcm_sample_bit_depth_chroma_minus1);
+	GET_HEVC_VDEC_PARAM(log2_min_pcm_luma_coding_block_size_minus3);
+	GET_HEVC_VDEC_PARAM(log2_diff_max_min_pcm_luma_coding_block_size);
+	GET_HEVC_VDEC_PARAM(num_short_term_ref_pic_sets);
+	GET_HEVC_VDEC_PARAM(num_long_term_ref_pics_sps);
+	GET_HEVC_VDEC_PARAM(chroma_format_idc);
+	GET_HEVC_VDEC_PARAM(sps_max_sub_layers_minus1);
+
+	GET_HEVC_VDEC_FLAG(separate_colour_plane,
+			   V4L2_HEVC_SPS_FLAG_SEPARATE_COLOUR_PLANE);
+	GET_HEVC_VDEC_FLAG(scaling_list_enabled,
+			   V4L2_HEVC_SPS_FLAG_SCALING_LIST_ENABLED);
+	GET_HEVC_VDEC_FLAG(amp_enabled,
+			   V4L2_HEVC_SPS_FLAG_AMP_ENABLED);
+	GET_HEVC_VDEC_FLAG(sample_adaptive_offset,
+			   V4L2_HEVC_SPS_FLAG_SAMPLE_ADAPTIVE_OFFSET);
+	GET_HEVC_VDEC_FLAG(pcm_enabled,
+			   V4L2_HEVC_SPS_FLAG_PCM_ENABLED);
+	GET_HEVC_VDEC_FLAG(pcm_loop_filter_disabled,
+			   V4L2_HEVC_SPS_FLAG_PCM_LOOP_FILTER_DISABLED);
+	GET_HEVC_VDEC_FLAG(long_term_ref_pics_enabled,
+			   V4L2_HEVC_SPS_FLAG_LONG_TERM_REF_PICS_PRESENT);
+	GET_HEVC_VDEC_FLAG(sps_temporal_mvp_enabled,
+			   V4L2_HEVC_SPS_FLAG_SPS_TEMPORAL_MVP_ENABLED);
+	GET_HEVC_VDEC_FLAG(strong_intra_smoothing_enabled,
+			   V4L2_HEVC_SPS_FLAG_STRONG_INTRA_SMOOTHING_ENABLED);
+}
+
+static void vdec_hevc_copy_pps_params(struct mtk_hevc_pps_param *dst_param,
+				      const struct v4l2_ctrl_hevc_pps *src_param)
+{
+	int i;
+
+	GET_HEVC_VDEC_PARAM(pic_parameter_set_id);
+	GET_HEVC_VDEC_PARAM(num_extra_slice_header_bits);
+	GET_HEVC_VDEC_PARAM(num_ref_idx_l0_default_active_minus1);
+	GET_HEVC_VDEC_PARAM(num_ref_idx_l1_default_active_minus1);
+	GET_HEVC_VDEC_PARAM(init_qp_minus26);
+	GET_HEVC_VDEC_PARAM(diff_cu_qp_delta_depth);
+	GET_HEVC_VDEC_PARAM(pps_cb_qp_offset);
+	GET_HEVC_VDEC_PARAM(pps_cr_qp_offset);
+	GET_HEVC_VDEC_PARAM(num_tile_columns_minus1);
+	GET_HEVC_VDEC_PARAM(num_tile_rows_minus1);
+	GET_HEVC_VDEC_PARAM(init_qp_minus26);
+	GET_HEVC_VDEC_PARAM(diff_cu_qp_delta_depth);
+	GET_HEVC_VDEC_PARAM(pic_parameter_set_id);
+	GET_HEVC_VDEC_PARAM(num_extra_slice_header_bits);
+	GET_HEVC_VDEC_PARAM(num_ref_idx_l0_default_active_minus1);
+	GET_HEVC_VDEC_PARAM(num_ref_idx_l1_default_active_minus1);
+	GET_HEVC_VDEC_PARAM(pps_beta_offset_div2);
+	GET_HEVC_VDEC_PARAM(pps_tc_offset_div2);
+	GET_HEVC_VDEC_PARAM(log2_parallel_merge_level_minus2);
+
+	for (i = 0; i < ARRAY_SIZE(src_param->column_width_minus1); i++)
+		GET_HEVC_VDEC_PARAM(column_width_minus1[i]);
+	for (i = 0; i < ARRAY_SIZE(src_param->row_height_minus1); i++)
+		GET_HEVC_VDEC_PARAM(row_height_minus1[i]);
+
+	GET_HEVC_VDEC_FLAG(dependent_slice_segment_enabled,
+			   V4L2_HEVC_PPS_FLAG_DEPENDENT_SLICE_SEGMENT_ENABLED);
+	GET_HEVC_VDEC_FLAG(output_flag_present,
+			   V4L2_HEVC_PPS_FLAG_OUTPUT_FLAG_PRESENT);
+	GET_HEVC_VDEC_FLAG(sign_data_hiding_enabled,
+			   V4L2_HEVC_PPS_FLAG_SIGN_DATA_HIDING_ENABLED);
+	GET_HEVC_VDEC_FLAG(cabac_init_present,
+			   V4L2_HEVC_PPS_FLAG_CABAC_INIT_PRESENT);
+	GET_HEVC_VDEC_FLAG(constrained_intra_pred,
+			   V4L2_HEVC_PPS_FLAG_CONSTRAINED_INTRA_PRED);
+	GET_HEVC_VDEC_FLAG(transform_skip_enabled,
+			   V4L2_HEVC_PPS_FLAG_TRANSFORM_SKIP_ENABLED);
+	GET_HEVC_VDEC_FLAG(cu_qp_delta_enabled,
+			   V4L2_HEVC_PPS_FLAG_CU_QP_DELTA_ENABLED);
+	GET_HEVC_VDEC_FLAG(pps_slice_chroma_qp_offsets_present,
+			   V4L2_HEVC_PPS_FLAG_PPS_SLICE_CHROMA_QP_OFFSETS_PRESENT);
+	GET_HEVC_VDEC_FLAG(weighted_pred,
+			   V4L2_HEVC_PPS_FLAG_WEIGHTED_PRED);
+	GET_HEVC_VDEC_FLAG(weighted_bipred,
+			   V4L2_HEVC_PPS_FLAG_WEIGHTED_BIPRED);
+	GET_HEVC_VDEC_FLAG(transquant_bypass_enabled,
+			   V4L2_HEVC_PPS_FLAG_TRANSQUANT_BYPASS_ENABLED);
+	GET_HEVC_VDEC_FLAG(pps_flag_tiles_enabled,
+			   V4L2_HEVC_PPS_FLAG_TILES_ENABLED);
+	GET_HEVC_VDEC_FLAG(entropy_coding_sync_enabled,
+			   V4L2_HEVC_PPS_FLAG_ENTROPY_CODING_SYNC_ENABLED);
+	GET_HEVC_VDEC_FLAG(loop_filter_across_tiles_enabled,
+			   V4L2_HEVC_PPS_FLAG_LOOP_FILTER_ACROSS_TILES_ENABLED);
+	GET_HEVC_VDEC_FLAG(pps_loop_filter_across_slices_enabled,
+			   V4L2_HEVC_PPS_FLAG_PPS_LOOP_FILTER_ACROSS_SLICES_ENABLED);
+	GET_HEVC_VDEC_FLAG(deblocking_filter_override_enabled,
+			   V4L2_HEVC_PPS_FLAG_DEBLOCKING_FILTER_OVERRIDE_ENABLED);
+	GET_HEVC_VDEC_FLAG(pps_disable_deflocking_filter,
+			   V4L2_HEVC_PPS_FLAG_PPS_DISABLE_DEBLOCKING_FILTER);
+	GET_HEVC_VDEC_FLAG(lists_modification_present,
+			   V4L2_HEVC_PPS_FLAG_LISTS_MODIFICATION_PRESENT);
+	GET_HEVC_VDEC_FLAG(slice_segment_header_extersion_present,
+			   V4L2_HEVC_PPS_FLAG_SLICE_SEGMENT_HEADER_EXTENSION_PRESENT);
+	GET_HEVC_VDEC_FLAG(deblocking_filter_control_present,
+			   V4L2_HEVC_PPS_FLAG_DEBLOCKING_FILTER_CONTROL_PRESENT);
+	GET_HEVC_VDEC_FLAG(uniform_spacing,
+			   V4L2_HEVC_PPS_FLAG_UNIFORM_SPACING);
+}
+
+static void vdec_hevc_copy_scaling_matrix(struct slice_api_hevc_scaling_matrix *dst_matrix,
+					  const struct v4l2_ctrl_hevc_scaling_matrix *src_matrix)
+{
+	memcpy(dst_matrix, src_matrix, sizeof(*src_matrix));
+}
+
+static void
+vdec_hevc_copy_decode_params(struct slice_api_hevc_decode_param *dst_param,
+			     const struct v4l2_ctrl_hevc_decode_params *src_param,
+			     const struct v4l2_hevc_dpb_entry dpb[V4L2_HEVC_DPB_ENTRIES_NUM_MAX])
+{
+	struct slice_hevc_dpb_entry *dst_entry;
+	const struct v4l2_hevc_dpb_entry *src_entry;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(dst_param->dpb); i++) {
+		dst_entry = &dst_param->dpb[i];
+		src_entry = &dpb[i];
+
+		dst_entry->timestamp = src_entry->timestamp;
+		dst_entry->flags = src_entry->flags;
+		dst_entry->field_pic = src_entry->field_pic;
+		dst_entry->pic_order_cnt_val = src_entry->pic_order_cnt_val;
+
+		GET_HEVC_VDEC_PARAM(poc_st_curr_before[i]);
+		GET_HEVC_VDEC_PARAM(poc_st_curr_after[i]);
+		GET_HEVC_VDEC_PARAM(poc_lt_curr[i]);
+	}
+
+	GET_HEVC_VDEC_PARAM(pic_order_cnt_val);
+	GET_HEVC_VDEC_PARAM(short_term_ref_pic_set_size);
+	GET_HEVC_VDEC_PARAM(long_term_ref_pic_set_size);
+	GET_HEVC_VDEC_PARAM(num_active_dpb_entries);
+	GET_HEVC_VDEC_PARAM(num_poc_st_curr_before);
+	GET_HEVC_VDEC_PARAM(num_poc_st_curr_after);
+	GET_HEVC_VDEC_PARAM(num_delta_pocs_of_ref_rps_idx);
+	GET_HEVC_VDEC_PARAM(num_poc_lt_curr);
+	GET_HEVC_VDEC_PARAM(flags);
+}
+
+static int vdec_hevc_slice_fill_decode_parameters(struct vdec_hevc_slice_inst *inst,
+						  struct vdec_hevc_slice_share_info *share_info)
+{
+	struct vdec_hevc_slice_lat_dec_param *slice_param = &inst->vsi->hevc_slice_params;
+	const struct v4l2_ctrl_hevc_decode_params *dec_params;
+	const struct v4l2_ctrl_hevc_scaling_matrix *src_matrix;
+	const struct v4l2_ctrl_hevc_sps *sps;
+	const struct v4l2_ctrl_hevc_pps *pps;
+
+	dec_params =
+		vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_DECODE_PARAMS);
+	if (IS_ERR(dec_params))
+		return PTR_ERR(dec_params);
+
+	src_matrix =
+		vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_SCALING_MATRIX);
+	if (IS_ERR(src_matrix))
+		return PTR_ERR(src_matrix);
+
+	sps = vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_SPS);
+	if (IS_ERR(sps))
+		return PTR_ERR(sps);
+
+	pps = vdec_hevc_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_HEVC_PPS);
+	if (IS_ERR(pps))
+		return PTR_ERR(pps);
+
+	vdec_hevc_copy_sps_params(&slice_param->sps, sps);
+	vdec_hevc_copy_pps_params(&slice_param->pps, pps);
+	vdec_hevc_copy_scaling_matrix(&slice_param->scaling_matrix, src_matrix);
+
+	memcpy(&share_info->sps, sps, sizeof(*sps));
+	memcpy(&share_info->dec_params, dec_params, sizeof(*dec_params));
+
+	slice_param->decode_params.num_poc_st_curr_before = dec_params->num_poc_st_curr_before;
+	slice_param->decode_params.num_poc_st_curr_after = dec_params->num_poc_st_curr_after;
+	slice_param->decode_params.num_poc_lt_curr = dec_params->num_poc_lt_curr;
+	slice_param->decode_params.num_delta_pocs_of_ref_rps_idx =
+		dec_params->num_delta_pocs_of_ref_rps_idx;
+
+	return 0;
+}
+
+static void vdec_hevc_slice_fill_decode_reflist(struct vdec_hevc_slice_inst *inst,
+						struct vdec_hevc_slice_lat_dec_param *slice_param,
+						struct vdec_hevc_slice_share_info *share_info)
+{
+	struct v4l2_ctrl_hevc_decode_params *dec_params = &share_info->dec_params;
+
+	vdec_hevc_copy_decode_params(&slice_param->decode_params, dec_params,
+				     share_info->dec_params.dpb);
+
+	vdec_hevc_fill_dpb_info(inst->ctx, &slice_param->decode_params,
+				slice_param->hevc_dpb_info);
+}
+
+static int vdec_hevc_slice_alloc_mv_buf(struct vdec_hevc_slice_inst *inst,
+					struct vdec_pic_info *pic)
+{
+	unsigned int buf_sz = vdec_hevc_get_mv_buf_size(pic->buf_w, pic->buf_h);
+	struct mtk_vcodec_mem *mem;
+	int i, err;
+
+	mtk_v4l2_vdec_dbg(3, inst->ctx, "allocate mv buffer size = 0x%x", buf_sz);
+	for (i = 0; i < HEVC_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+		mem->size = buf_sz;
+		err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+		if (err) {
+			mtk_vdec_err(inst->ctx, "failed to allocate mv buf");
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static void vdec_hevc_slice_free_mv_buf(struct vdec_hevc_slice_inst *inst)
+{
+	int i;
+	struct mtk_vcodec_mem *mem;
+
+	for (i = 0; i < HEVC_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		if (mem->va)
+			mtk_vcodec_mem_free(inst->ctx, mem);
+	}
+}
+
+static void vdec_hevc_slice_get_pic_info(struct vdec_hevc_slice_inst *inst)
+{
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+	u32 data[3];
+
+	data[0] = ctx->picinfo.pic_w;
+	data[1] = ctx->picinfo.pic_h;
+	data[2] = ctx->capture_fourcc;
+	vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO);
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, VCODEC_DEC_ALIGNED_64);
+	ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0];
+	ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1];
+	inst->cap_num_planes =
+		ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes;
+
+	mtk_vdec_debug(ctx, "pic(%d, %d), buf(%d, %d)",
+		       ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+		       ctx->picinfo.buf_w, ctx->picinfo.buf_h);
+	mtk_vdec_debug(ctx, "Y/C(%d, %d)", ctx->picinfo.fb_sz[0],
+		       ctx->picinfo.fb_sz[1]);
+
+	if (ctx->last_decoded_picinfo.pic_w != ctx->picinfo.pic_w ||
+	    ctx->last_decoded_picinfo.pic_h != ctx->picinfo.pic_h) {
+		inst->resolution_changed = true;
+		if (ctx->last_decoded_picinfo.buf_w != ctx->picinfo.buf_w ||
+		    ctx->last_decoded_picinfo.buf_h != ctx->picinfo.buf_h)
+			inst->realloc_mv_buf = true;
+
+		mtk_v4l2_vdec_dbg(1, inst->ctx, "resChg: (%d %d) : old(%d, %d) -> new(%d, %d)",
+				  inst->resolution_changed,
+				  inst->realloc_mv_buf,
+				  ctx->last_decoded_picinfo.pic_w,
+				  ctx->last_decoded_picinfo.pic_h,
+				  ctx->picinfo.pic_w, ctx->picinfo.pic_h);
+	}
+}
+
+static void vdec_hevc_slice_get_crop_info(struct vdec_hevc_slice_inst *inst,
+					  struct v4l2_rect *cr)
+{
+	cr->left = 0;
+	cr->top = 0;
+	cr->width = inst->ctx->picinfo.pic_w;
+	cr->height = inst->ctx->picinfo.pic_h;
+
+	mtk_vdec_debug(inst->ctx, "l=%d, t=%d, w=%d, h=%d",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static int vdec_hevc_slice_setup_lat_buffer(struct vdec_hevc_slice_inst *inst,
+					    struct mtk_vcodec_mem *bs,
+					    struct vdec_lat_buf *lat_buf,
+					    bool *res_chg)
+{
+	struct mtk_vcodec_mem *mem;
+	struct mtk_video_dec_buf *src_buf_info;
+	struct vdec_hevc_slice_share_info *share_info;
+	int i, err;
+
+	inst->vsi->bs.dma_addr = (u64)bs->dma_addr;
+	inst->vsi->bs.size = bs->size;
+
+	src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer);
+	lat_buf->src_buf_req = src_buf_info->m2m_buf.vb.vb2_buf.req_obj.req;
+	v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb, &lat_buf->ts_info, true);
+
+	*res_chg = inst->resolution_changed;
+	if (inst->resolution_changed) {
+		mtk_vdec_debug(inst->ctx, "- resolution changed -");
+		if (inst->realloc_mv_buf) {
+			err = vdec_hevc_slice_alloc_mv_buf(inst, &inst->ctx->picinfo);
+			inst->realloc_mv_buf = false;
+			if (err)
+				return err;
+		}
+		inst->resolution_changed = false;
+	}
+
+	for (i = 0; i < HEVC_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		inst->vsi->mv_buf_dma[i].dma_addr = mem->dma_addr;
+		inst->vsi->mv_buf_dma[i].size = mem->size;
+	}
+
+	inst->vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	inst->vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size;
+
+	inst->vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr;
+	inst->vsi->err_map.size = lat_buf->wdma_err_addr.size;
+
+	inst->vsi->slice_bc.dma_addr = lat_buf->slice_bc_addr.dma_addr;
+	inst->vsi->slice_bc.size = lat_buf->slice_bc_addr.size;
+
+	inst->vsi->trans.dma_addr_end = inst->ctx->msg_queue.wdma_rptr_addr;
+	inst->vsi->trans.dma_addr = inst->ctx->msg_queue.wdma_wptr_addr;
+
+	share_info = lat_buf->private_data;
+	share_info->trans.dma_addr = inst->vsi->trans.dma_addr;
+	share_info->trans.dma_addr_end = inst->vsi->trans.dma_addr_end;
+
+	mtk_vdec_debug(inst->ctx, "lat: ube addr/size(0x%llx 0x%llx) err:0x%llx",
+		       inst->vsi->ube.buf,
+		       inst->vsi->ube.padding,
+		       inst->vsi->err_map.buf);
+
+	mtk_vdec_debug(inst->ctx, "slice addr/size(0x%llx 0x%llx) trans start/end((0x%llx 0x%llx))",
+		       inst->vsi->slice_bc.buf,
+		       inst->vsi->slice_bc.padding,
+		       inst->vsi->trans.buf,
+		       inst->vsi->trans.padding);
+
+	return 0;
+}
+
+static int vdec_hevc_slice_setup_core_buffer(struct vdec_hevc_slice_inst *inst,
+					     struct vdec_hevc_slice_share_info *share_info,
+					     struct vdec_lat_buf *lat_buf)
+{
+	struct mtk_vcodec_mem *mem;
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+	struct vb2_v4l2_buffer *vb2_v4l2;
+	struct vdec_fb *fb;
+	u64 y_fb_dma, c_fb_dma;
+	int i;
+
+	fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx);
+	if (!fb) {
+		mtk_vdec_err(inst->ctx, "fb buffer is NULL");
+		return -EBUSY;
+	}
+
+	y_fb_dma = (u64)fb->base_y.dma_addr;
+	if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1)
+		c_fb_dma =
+			y_fb_dma + inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h;
+	else
+		c_fb_dma = (u64)fb->base_c.dma_addr;
+
+	mtk_vdec_debug(inst->ctx, "[hevc-core] y/c addr = 0x%llx 0x%llx", y_fb_dma, c_fb_dma);
+
+	inst->vsi_core->fb.y.dma_addr = y_fb_dma;
+	inst->vsi_core->fb.y.size = ctx->picinfo.fb_sz[0];
+	inst->vsi_core->fb.c.dma_addr = c_fb_dma;
+	inst->vsi_core->fb.y.size = ctx->picinfo.fb_sz[1];
+
+	inst->vsi_core->dec.vdec_fb_va = (unsigned long)fb;
+
+	inst->vsi_core->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	inst->vsi_core->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size;
+
+	inst->vsi_core->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr;
+	inst->vsi_core->err_map.size = lat_buf->wdma_err_addr.size;
+
+	inst->vsi_core->slice_bc.dma_addr = lat_buf->slice_bc_addr.dma_addr;
+	inst->vsi_core->slice_bc.size = lat_buf->slice_bc_addr.size;
+
+	inst->vsi_core->trans.dma_addr = share_info->trans.dma_addr;
+	inst->vsi_core->trans.dma_addr_end = share_info->trans.dma_addr_end;
+
+	inst->vsi_core->wrap.dma_addr = inst->wrap_addr.dma_addr;
+	inst->vsi_core->wrap.size = inst->wrap_addr.size;
+
+	for (i = 0; i < HEVC_MAX_MV_NUM; i++) {
+		mem = &inst->mv_buf[i];
+		inst->vsi_core->mv_buf_dma[i].dma_addr = mem->dma_addr;
+		inst->vsi_core->mv_buf_dma[i].size = mem->size;
+	}
+
+	vb2_v4l2 = v4l2_m2m_next_dst_buf(ctx->m2m_ctx);
+	v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, vb2_v4l2, true);
+
+	return 0;
+}
+
+static int vdec_hevc_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_hevc_slice_inst *inst;
+	int err, vsi_size;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = SCP_IPI_VDEC_LAT;
+	inst->vpu.core_id = SCP_IPI_VDEC_CORE;
+	inst->vpu.ctx = ctx;
+	inst->vpu.codec_type = ctx->current_codec;
+	inst->vpu.capture_type = ctx->capture_fourcc;
+
+	ctx->drv_handle = inst;
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_hevc init err=%d", err);
+		goto error_free_inst;
+	}
+
+	vsi_size = round_up(sizeof(struct vdec_hevc_slice_vsi), VCODEC_DEC_ALIGNED_64);
+	inst->vsi = inst->vpu.vsi;
+	inst->vsi_core =
+		(struct vdec_hevc_slice_vsi *)(((char *)inst->vpu.vsi) + vsi_size);
+
+	inst->resolution_changed = true;
+	inst->realloc_mv_buf = true;
+
+	inst->wrap_addr.size = VDEC_HEVC_WRAP_SZ;
+	err = mtk_vcodec_mem_alloc(ctx, &inst->wrap_addr);
+	if (err)
+		goto error_free_inst;
+
+	mtk_vdec_debug(ctx, "lat struct size = %d,%d,%d,%d vsi: %d\n",
+		       (int)sizeof(struct mtk_hevc_sps_param),
+		       (int)sizeof(struct mtk_hevc_pps_param),
+		       (int)sizeof(struct vdec_hevc_slice_lat_dec_param),
+		       (int)sizeof(struct mtk_hevc_dpb_info),
+			 vsi_size);
+	mtk_vdec_debug(ctx, "lat hevc instance >> %p, codec_type = 0x%x",
+		       inst, inst->vpu.codec_type);
+
+	return 0;
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static void vdec_hevc_slice_deinit(void *h_vdec)
+{
+	struct vdec_hevc_slice_inst *inst = h_vdec;
+	struct mtk_vcodec_mem *mem;
+
+	vpu_dec_deinit(&inst->vpu);
+	vdec_hevc_slice_free_mv_buf(inst);
+
+	mem = &inst->wrap_addr;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	vdec_msg_queue_deinit(&inst->ctx->msg_queue, inst->ctx);
+	kfree(inst);
+}
+
+static int vdec_hevc_slice_core_decode(struct vdec_lat_buf *lat_buf)
+{
+	int err, timeout;
+	struct mtk_vcodec_dec_ctx *ctx = lat_buf->ctx;
+	struct vdec_hevc_slice_inst *inst = ctx->drv_handle;
+	struct vdec_hevc_slice_share_info *share_info = lat_buf->private_data;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+
+	mtk_vdec_debug(ctx, "[hevc-core] vdec_hevc core decode");
+	memcpy(&inst->vsi_core->hevc_slice_params, &share_info->hevc_slice_params,
+	       sizeof(share_info->hevc_slice_params));
+
+	err = vdec_hevc_slice_setup_core_buffer(inst, share_info, lat_buf);
+	if (err)
+		goto vdec_dec_end;
+
+	vdec_hevc_slice_fill_decode_reflist(inst, &inst->vsi_core->hevc_slice_params,
+					    share_info);
+	err = vpu_dec_core(vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "core decode err=%d", err);
+		goto vdec_dec_end;
+	}
+
+	/* wait decoder done interrupt */
+	timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					       WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE);
+	if (timeout)
+		mtk_vdec_err(ctx, "core decode timeout: pic_%d", ctx->decoded_frame_cnt);
+	inst->vsi_core->dec.timeout = !!timeout;
+
+	vpu_dec_core_end(vpu);
+	mtk_vdec_debug(ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x",
+		       ctx->decoded_frame_cnt,
+		       inst->vsi_core->dec.crc[0], inst->vsi_core->dec.crc[1],
+		       inst->vsi_core->dec.crc[2], inst->vsi_core->dec.crc[3],
+		       inst->vsi_core->dec.crc[4], inst->vsi_core->dec.crc[5],
+		       inst->vsi_core->dec.crc[6], inst->vsi_core->dec.crc[7]);
+
+vdec_dec_end:
+	vdec_msg_queue_update_ube_rptr(&lat_buf->ctx->msg_queue, share_info->trans.dma_addr_end);
+	ctx->dev->vdec_pdata->cap_to_disp(ctx, !!err, lat_buf->src_buf_req);
+	mtk_vdec_debug(ctx, "core decode done err=%d", err);
+	ctx->decoded_frame_cnt++;
+	return 0;
+}
+
+static int vdec_hevc_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				      struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_hevc_slice_inst *inst = h_vdec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	int err, timeout = 0;
+	unsigned int data[2];
+	struct vdec_lat_buf *lat_buf;
+	struct vdec_hevc_slice_share_info *share_info;
+
+	if (vdec_msg_queue_init(&inst->ctx->msg_queue, inst->ctx,
+				vdec_hevc_slice_core_decode,
+				sizeof(*share_info)))
+		return -ENOMEM;
+
+	/* bs NULL means flush decoder */
+	if (!bs) {
+		vdec_msg_queue_wait_lat_buf_full(&inst->ctx->msg_queue);
+		return vpu_dec_reset(vpu);
+	}
+
+	lat_buf = vdec_msg_queue_dqbuf(&inst->ctx->msg_queue.lat_ctx);
+	if (!lat_buf) {
+		mtk_vdec_debug(inst->ctx, "failed to get lat buffer");
+		return -EAGAIN;
+	}
+
+	share_info = lat_buf->private_data;
+	err = vdec_hevc_slice_fill_decode_parameters(inst, share_info);
+	if (err)
+		goto err_free_fb_out;
+
+	err = vdec_hevc_slice_setup_lat_buffer(inst, bs, lat_buf, res_chg);
+	if (err)
+		goto err_free_fb_out;
+
+	err = vpu_dec_start(vpu, data, 2);
+	if (err) {
+		mtk_vdec_debug(inst->ctx, "lat decode err: %d", err);
+		goto err_free_fb_out;
+	}
+
+	if (IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) {
+		memcpy(&share_info->hevc_slice_params, &inst->vsi->hevc_slice_params,
+		       sizeof(share_info->hevc_slice_params));
+		vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf);
+	}
+
+	/* wait decoder done interrupt */
+	timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					       WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0);
+	if (timeout)
+		mtk_vdec_err(inst->ctx, "lat decode timeout: pic_%d", inst->slice_dec_num);
+	inst->vsi->dec.timeout = !!timeout;
+
+	err = vpu_dec_end(vpu);
+	if (err == SLICE_HEADER_FULL || err == TRANS_BUFFER_FULL) {
+		if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability))
+			vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf);
+		inst->slice_dec_num++;
+		mtk_vdec_err(inst->ctx, "lat dec fail: pic_%d err:%d", inst->slice_dec_num, err);
+		return -EINVAL;
+	}
+
+	share_info->trans.dma_addr_end = inst->ctx->msg_queue.wdma_addr.dma_addr +
+		inst->vsi->dec.wdma_end_addr_offset;
+	vdec_msg_queue_update_ube_wptr(&lat_buf->ctx->msg_queue, share_info->trans.dma_addr_end);
+
+	if (!IS_VDEC_INNER_RACING(inst->ctx->dev->dec_capability)) {
+		memcpy(&share_info->hevc_slice_params, &inst->vsi->hevc_slice_params,
+		       sizeof(share_info->hevc_slice_params));
+		vdec_msg_queue_qbuf(&inst->ctx->msg_queue.core_ctx, lat_buf);
+	}
+	mtk_vdec_debug(inst->ctx, "dec num: %d lat crc: 0x%x 0x%x 0x%x", inst->slice_dec_num,
+		       inst->vsi->dec.crc[0], inst->vsi->dec.crc[1], inst->vsi->dec.crc[2]);
+
+	inst->slice_dec_num++;
+	return 0;
+err_free_fb_out:
+	vdec_msg_queue_qbuf(&inst->ctx->msg_queue.lat_ctx, lat_buf);
+	mtk_vdec_err(inst->ctx, "slice dec number: %d err: %d", inst->slice_dec_num, err);
+	return err;
+}
+
+static int vdec_hevc_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				  struct vdec_fb *unused, bool *res_chg)
+{
+	struct vdec_hevc_slice_inst *inst = h_vdec;
+
+	if (!h_vdec || inst->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE)
+		return -EINVAL;
+
+	return vdec_hevc_slice_lat_decode(h_vdec, bs, unused, res_chg);
+}
+
+static int vdec_hevc_slice_get_param(void *h_vdec, enum vdec_get_param_type type,
+				     void *out)
+{
+	struct vdec_hevc_slice_inst *inst = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		vdec_hevc_slice_get_pic_info(inst);
+		break;
+	case GET_PARAM_DPB_SIZE:
+		*(unsigned int *)out = 6;
+		break;
+	case GET_PARAM_CROP_INFO:
+		vdec_hevc_slice_get_crop_info(inst, out);
+		break;
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+	return 0;
+}
+
+const struct vdec_common_if vdec_hevc_slice_multi_if = {
+	.init		= vdec_hevc_slice_init,
+	.decode		= vdec_hevc_slice_decode,
+	.get_param	= vdec_hevc_slice_get_param,
+	.deinit		= vdec_hevc_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c
new file mode 100644
index 0000000000..19407f9bc7
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c
@@ -0,0 +1,608 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: Jungchang Tsao <jungchang.tsao@mediatek.com>
+ *	   PC Chen <pc.chen@mediatek.com>
+ */
+
+#include <linux/slab.h>
+#include "../vdec_drv_if.h"
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_vpu_if.h"
+#include "../vdec_drv_base.h"
+
+/* Decoding picture buffer size (3 reference frames plus current frame) */
+#define VP8_DPB_SIZE			4
+
+/* HW working buffer size (bytes) */
+#define VP8_WORKING_BUF_SZ		(45 * 4096)
+
+/* HW control register address */
+#define VP8_SEGID_DRAM_ADDR		0x3c
+#define VP8_HW_VLD_ADDR			0x93C
+#define VP8_HW_VLD_VALUE		0x940
+#define VP8_BSASET			0x100
+#define VP8_BSDSET			0x104
+#define VP8_RW_CKEN_SET			0x0
+#define VP8_RW_DCM_CON			0x18
+#define VP8_WO_VLD_SRST			0x108
+#define VP8_RW_MISC_SYS_SEL		0x84
+#define VP8_RW_MISC_SPEC_CON		0xC8
+#define VP8_WO_VLD_SRST			0x108
+#define VP8_RW_VP8_CTRL			0xA4
+#define VP8_RW_MISC_DCM_CON		0xEC
+#define VP8_RW_MISC_SRST		0xF4
+#define VP8_RW_MISC_FUNC_CON		0xCC
+
+#define VP8_MAX_FRM_BUF_NUM		5
+#define VP8_MAX_FRM_BUF_NODE_NUM	(VP8_MAX_FRM_BUF_NUM * 2)
+
+/* required buffer size (bytes) to store decode information */
+#define VP8_HW_SEGMENT_DATA_SZ		272
+#define VP8_HW_SEGMENT_UINT		4
+
+#define VP8_DEC_TABLE_PROC_LOOP		96
+#define VP8_DEC_TABLE_UNIT		3
+#define VP8_DEC_TABLE_SZ		300
+#define VP8_DEC_TABLE_OFFSET		2
+#define VP8_DEC_TABLE_RW_UNIT		4
+
+/**
+ * struct vdec_vp8_dec_info - decode misc information
+ * @working_buf_dma   : working buffer dma address
+ * @prev_y_dma        : previous decoded frame buffer Y plane address
+ * @cur_y_fb_dma      : current plane Y frame buffer dma address
+ * @cur_c_fb_dma      : current plane C frame buffer dma address
+ * @bs_dma	      : bitstream dma address
+ * @bs_sz	      : bitstream size
+ * @resolution_changed: resolution change flag 1 - changed,  0 - not change
+ * @show_frame	      : display this frame or not
+ * @wait_key_frame    : wait key frame coming
+ */
+struct vdec_vp8_dec_info {
+	uint64_t working_buf_dma;
+	uint64_t prev_y_dma;
+	uint64_t cur_y_fb_dma;
+	uint64_t cur_c_fb_dma;
+	uint64_t bs_dma;
+	uint32_t bs_sz;
+	uint32_t resolution_changed;
+	uint32_t show_frame;
+	uint32_t wait_key_frame;
+};
+
+/**
+ * struct vdec_vp8_vsi - VPU shared information
+ * @dec			: decoding information
+ * @pic			: picture information
+ * @dec_table		: decoder coefficient table
+ * @segment_buf		: segmentation buffer
+ * @load_data		: flag to indicate reload decode data
+ */
+struct vdec_vp8_vsi {
+	struct vdec_vp8_dec_info dec;
+	struct vdec_pic_info pic;
+	uint32_t dec_table[VP8_DEC_TABLE_SZ];
+	uint32_t segment_buf[VP8_HW_SEGMENT_DATA_SZ][VP8_HW_SEGMENT_UINT];
+	uint32_t load_data;
+};
+
+/**
+ * struct vdec_vp8_hw_reg_base - HW register base
+ * @misc	: base address for misc
+ * @ld		: base address for ld
+ * @top		: base address for top
+ * @cm		: base address for cm
+ * @hwd		: base address for hwd
+ * @hwb		: base address for hwb
+ */
+struct vdec_vp8_hw_reg_base {
+	void __iomem *misc;
+	void __iomem *ld;
+	void __iomem *top;
+	void __iomem *cm;
+	void __iomem *hwd;
+	void __iomem *hwb;
+};
+
+/**
+ * struct vdec_vp8_vpu_inst - VPU instance for VP8 decode
+ * @wq_hd	: Wait queue to wait VPU message ack
+ * @signaled	: 1 - Host has received ack message from VPU, 0 - not receive
+ * @failure	: VPU execution result status 0 - success, others - fail
+ * @inst_addr	: VPU decoder instance address
+ */
+struct vdec_vp8_vpu_inst {
+	wait_queue_head_t wq_hd;
+	int signaled;
+	int failure;
+	uint32_t inst_addr;
+};
+
+/* frame buffer (fb) list
+ * [available_fb_node_list]  - decode fb are initialized to 0 and populated in
+ * [fb_use_list]  - fb is set after decode and is moved to this list
+ * [fb_free_list] - fb is not needed for reference will be moved from
+ *		     [fb_use_list] to [fb_free_list] and
+ *		     once user remove fb from [fb_free_list],
+ *		     it is circulated back to [available_fb_node_list]
+ * [fb_disp_list] - fb is set after decode and is moved to this list
+ *                   once user remove fb from [fb_disp_list] it is
+ *                   circulated back to [available_fb_node_list]
+ */
+
+/**
+ * struct vdec_vp8_inst - VP8 decoder instance
+ * @cur_fb		   : current frame buffer
+ * @dec_fb		   : decode frame buffer node
+ * @available_fb_node_list : list to store available frame buffer node
+ * @fb_use_list		   : list to store frame buffer in use
+ * @fb_free_list	   : list to store free frame buffer
+ * @fb_disp_list	   : list to store display ready frame buffer
+ * @working_buf		   : HW decoder working buffer
+ * @reg_base		   : HW register base address
+ * @frm_cnt		   : decode frame count
+ * @ctx			   : V4L2 context
+ * @vpu			   : VPU instance for decoder
+ * @vsi			   : VPU share information
+ */
+struct vdec_vp8_inst {
+	struct vdec_fb *cur_fb;
+	struct vdec_fb_node dec_fb[VP8_MAX_FRM_BUF_NODE_NUM];
+	struct list_head available_fb_node_list;
+	struct list_head fb_use_list;
+	struct list_head fb_free_list;
+	struct list_head fb_disp_list;
+	struct mtk_vcodec_mem working_buf;
+	struct vdec_vp8_hw_reg_base reg_base;
+	unsigned int frm_cnt;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vpu_inst vpu;
+	struct vdec_vp8_vsi *vsi;
+};
+
+static void get_hw_reg_base(struct vdec_vp8_inst *inst)
+{
+	void __iomem **reg_base = inst->ctx->dev->reg_base;
+
+	inst->reg_base.top = mtk_vcodec_get_reg_addr(reg_base, VDEC_TOP);
+	inst->reg_base.cm = mtk_vcodec_get_reg_addr(reg_base, VDEC_CM);
+	inst->reg_base.hwd = mtk_vcodec_get_reg_addr(reg_base, VDEC_HWD);
+	inst->reg_base.misc = mtk_vcodec_get_reg_addr(reg_base, VDEC_MISC);
+	inst->reg_base.ld = mtk_vcodec_get_reg_addr(reg_base, VDEC_LD);
+	inst->reg_base.hwb = mtk_vcodec_get_reg_addr(reg_base, VDEC_HWB);
+}
+
+static void write_hw_segmentation_data(struct vdec_vp8_inst *inst)
+{
+	int i, j;
+	u32 seg_id_addr;
+	u32 val;
+	void __iomem *cm = inst->reg_base.cm;
+	struct vdec_vp8_vsi *vsi = inst->vsi;
+
+	seg_id_addr = readl(inst->reg_base.top + VP8_SEGID_DRAM_ADDR) >> 4;
+
+	for (i = 0; i < ARRAY_SIZE(vsi->segment_buf); i++) {
+		for (j = ARRAY_SIZE(vsi->segment_buf[i]) - 1; j >= 0; j--) {
+			val = (1 << 16) + ((seg_id_addr + i) << 2) + j;
+			writel(val, cm + VP8_HW_VLD_ADDR);
+
+			val = vsi->segment_buf[i][j];
+			writel(val, cm + VP8_HW_VLD_VALUE);
+		}
+	}
+}
+
+static void read_hw_segmentation_data(struct vdec_vp8_inst *inst)
+{
+	int i, j;
+	u32 seg_id_addr;
+	u32 val;
+	void __iomem *cm = inst->reg_base.cm;
+	struct vdec_vp8_vsi *vsi = inst->vsi;
+
+	seg_id_addr = readl(inst->reg_base.top + VP8_SEGID_DRAM_ADDR) >> 4;
+
+	for (i = 0; i < ARRAY_SIZE(vsi->segment_buf); i++) {
+		for (j = ARRAY_SIZE(vsi->segment_buf[i]) - 1; j >= 0; j--) {
+			val = ((seg_id_addr + i) << 2) + j;
+			writel(val, cm + VP8_HW_VLD_ADDR);
+
+			val = readl(cm + VP8_HW_VLD_VALUE);
+			vsi->segment_buf[i][j] = val;
+		}
+	}
+}
+
+/* reset HW and enable HW read/write data function */
+static void enable_hw_rw_function(struct vdec_vp8_inst *inst)
+{
+	u32 val = 0;
+	void __iomem *misc = inst->reg_base.misc;
+	void __iomem *ld = inst->reg_base.ld;
+	void __iomem *hwb = inst->reg_base.hwb;
+	void __iomem *hwd = inst->reg_base.hwd;
+
+	mtk_vcodec_write_vdecsys(inst->ctx, VP8_RW_CKEN_SET, 0x1);
+	writel(0x101, ld + VP8_WO_VLD_SRST);
+	writel(0x101, hwb + VP8_WO_VLD_SRST);
+
+	mtk_vcodec_write_vdecsys(inst->ctx, 0, 0x1);
+	val = readl(misc + VP8_RW_MISC_SRST);
+	writel((val & 0xFFFFFFFE), misc + VP8_RW_MISC_SRST);
+
+	writel(0x1, misc + VP8_RW_MISC_SYS_SEL);
+	writel(0x17F, misc + VP8_RW_MISC_SPEC_CON);
+	writel(0x71201100, misc + VP8_RW_MISC_FUNC_CON);
+	writel(0x0, ld + VP8_WO_VLD_SRST);
+	writel(0x0, hwb + VP8_WO_VLD_SRST);
+	mtk_vcodec_write_vdecsys(inst->ctx, VP8_RW_DCM_CON, 0x1);
+	writel(0x1, misc + VP8_RW_MISC_DCM_CON);
+	writel(0x1, hwd + VP8_RW_VP8_CTRL);
+}
+
+static void store_dec_table(struct vdec_vp8_inst *inst)
+{
+	int i, j;
+	u32 addr = 0, val = 0;
+	void __iomem *hwd = inst->reg_base.hwd;
+	u32 *p = &inst->vsi->dec_table[VP8_DEC_TABLE_OFFSET];
+
+	for (i = 0; i < VP8_DEC_TABLE_PROC_LOOP; i++) {
+		writel(addr, hwd + VP8_BSASET);
+		for (j = 0; j < VP8_DEC_TABLE_UNIT ; j++) {
+			val = *p++;
+			writel(val, hwd + VP8_BSDSET);
+		}
+		addr += VP8_DEC_TABLE_RW_UNIT;
+	}
+}
+
+static void load_dec_table(struct vdec_vp8_inst *inst)
+{
+	int i;
+	u32 addr = 0;
+	u32 *p = &inst->vsi->dec_table[VP8_DEC_TABLE_OFFSET];
+	void __iomem *hwd = inst->reg_base.hwd;
+
+	for (i = 0; i < VP8_DEC_TABLE_PROC_LOOP; i++) {
+		writel(addr, hwd + VP8_BSASET);
+		/* read total 11 bytes */
+		*p++ = readl(hwd + VP8_BSDSET);
+		*p++ = readl(hwd + VP8_BSDSET);
+		*p++ = readl(hwd + VP8_BSDSET) & 0xFFFFFF;
+		addr += VP8_DEC_TABLE_RW_UNIT;
+	}
+}
+
+static void get_pic_info(struct vdec_vp8_inst *inst, struct vdec_pic_info *pic)
+{
+	*pic = inst->vsi->pic;
+
+	mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)",
+		       pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h);
+	mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)",
+		       pic->fb_sz[0], pic->fb_sz[1]);
+}
+
+static void vp8_dec_finish(struct vdec_vp8_inst *inst)
+{
+	struct vdec_fb_node *node;
+	uint64_t prev_y_dma = inst->vsi->dec.prev_y_dma;
+
+	mtk_vdec_debug(inst->ctx, "prev fb base dma=%llx", prev_y_dma);
+
+	/* put last decode ok frame to fb_free_list */
+	if (prev_y_dma != 0) {
+		list_for_each_entry(node, &inst->fb_use_list, list) {
+			struct vdec_fb *fb = (struct vdec_fb *)node->fb;
+
+			if (prev_y_dma == (uint64_t)fb->base_y.dma_addr) {
+				list_move_tail(&node->list,
+					       &inst->fb_free_list);
+				break;
+			}
+		}
+	}
+
+	/* available_fb_node_list -> fb_use_list */
+	node = list_first_entry(&inst->available_fb_node_list,
+				struct vdec_fb_node, list);
+	node->fb = inst->cur_fb;
+	list_move_tail(&node->list, &inst->fb_use_list);
+
+	/* available_fb_node_list -> fb_disp_list */
+	if (inst->vsi->dec.show_frame) {
+		node = list_first_entry(&inst->available_fb_node_list,
+					struct vdec_fb_node, list);
+		node->fb = inst->cur_fb;
+		list_move_tail(&node->list, &inst->fb_disp_list);
+	}
+}
+
+static void move_fb_list_use_to_free(struct vdec_vp8_inst *inst)
+{
+	struct vdec_fb_node *node, *tmp;
+
+	list_for_each_entry_safe(node, tmp, &inst->fb_use_list, list)
+		list_move_tail(&node->list, &inst->fb_free_list);
+}
+
+static void init_list(struct vdec_vp8_inst *inst)
+{
+	int i;
+
+	INIT_LIST_HEAD(&inst->available_fb_node_list);
+	INIT_LIST_HEAD(&inst->fb_use_list);
+	INIT_LIST_HEAD(&inst->fb_free_list);
+	INIT_LIST_HEAD(&inst->fb_disp_list);
+
+	for (i = 0; i < ARRAY_SIZE(inst->dec_fb); i++) {
+		INIT_LIST_HEAD(&inst->dec_fb[i].list);
+		inst->dec_fb[i].fb = NULL;
+		list_add_tail(&inst->dec_fb[i].list,
+			      &inst->available_fb_node_list);
+	}
+}
+
+static void add_fb_to_free_list(struct vdec_vp8_inst *inst, void *fb)
+{
+	struct vdec_fb_node *node;
+
+	if (fb) {
+		node = list_first_entry(&inst->available_fb_node_list,
+					struct vdec_fb_node, list);
+		node->fb = fb;
+		list_move_tail(&node->list, &inst->fb_free_list);
+	}
+}
+
+static int alloc_working_buf(struct vdec_vp8_inst *inst)
+{
+	int err;
+	struct mtk_vcodec_mem *mem = &inst->working_buf;
+
+	mem->size = VP8_WORKING_BUF_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "Cannot allocate working buffer");
+		return err;
+	}
+
+	inst->vsi->dec.working_buf_dma = (uint64_t)mem->dma_addr;
+	return 0;
+}
+
+static void free_working_buf(struct vdec_vp8_inst *inst)
+{
+	struct mtk_vcodec_mem *mem = &inst->working_buf;
+
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	inst->vsi->dec.working_buf_dma = 0;
+}
+
+static int vdec_vp8_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_vp8_inst *inst;
+	int err;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return  -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = IPI_VDEC_VP8;
+	inst->vpu.ctx = ctx;
+
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_vp8 init err=%d", err);
+		goto error_free_inst;
+	}
+
+	inst->vsi = (struct vdec_vp8_vsi *)inst->vpu.vsi;
+	init_list(inst);
+	err = alloc_working_buf(inst);
+	if (err)
+		goto error_deinit;
+
+	get_hw_reg_base(inst);
+	mtk_vdec_debug(ctx, "VP8 Instance >> %p", inst);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+error_deinit:
+	vpu_dec_deinit(&inst->vpu);
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static int vdec_vp8_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+			   struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec;
+	struct vdec_vp8_dec_info *dec = &inst->vsi->dec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	unsigned char *bs_va;
+	unsigned int data;
+	int err = 0;
+	uint64_t y_fb_dma;
+	uint64_t c_fb_dma;
+
+	/* bs NULL means flush decoder */
+	if (bs == NULL) {
+		move_fb_list_use_to_free(inst);
+		return vpu_dec_reset(vpu);
+	}
+
+	y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0;
+	c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0;
+
+	mtk_vdec_debug(inst->ctx, "+ [%d] FB y_dma=%llx c_dma=%llx fb=%p",
+		       inst->frm_cnt, y_fb_dma, c_fb_dma, fb);
+
+	inst->cur_fb = fb;
+	dec->bs_dma = (unsigned long)bs->dma_addr;
+	dec->bs_sz = bs->size;
+	dec->cur_y_fb_dma = y_fb_dma;
+	dec->cur_c_fb_dma = c_fb_dma;
+
+	mtk_vdec_debug(inst->ctx, "\n + FRAME[%d] +\n", inst->frm_cnt);
+
+	write_hw_segmentation_data(inst);
+	enable_hw_rw_function(inst);
+	store_dec_table(inst);
+
+	bs_va = (unsigned char *)bs->va;
+
+	/* retrieve width/hight and scale info from header */
+	data = (*(bs_va + 9) << 24) | (*(bs_va + 8) << 16) |
+	       (*(bs_va + 7) << 8) | *(bs_va + 6);
+	err = vpu_dec_start(vpu, &data, 1);
+	if (err) {
+		add_fb_to_free_list(inst, fb);
+		if (dec->wait_key_frame) {
+			mtk_vdec_debug(inst->ctx, "wait key frame !");
+			return 0;
+		}
+
+		goto error;
+	}
+
+	if (dec->resolution_changed) {
+		mtk_vdec_debug(inst->ctx, "- resolution_changed -");
+		*res_chg = true;
+		add_fb_to_free_list(inst, fb);
+		return 0;
+	}
+
+	/* wait decoder done interrupt */
+	mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+				     WAIT_INTR_TIMEOUT_MS, 0);
+
+	if (inst->vsi->load_data)
+		load_dec_table(inst);
+
+	vp8_dec_finish(inst);
+	read_hw_segmentation_data(inst);
+
+	err = vpu_dec_end(vpu);
+	if (err)
+		goto error;
+
+	mtk_vdec_debug(inst->ctx, "\n - FRAME[%d] - show=%d\n", inst->frm_cnt, dec->show_frame);
+	inst->frm_cnt++;
+	*res_chg = false;
+	return 0;
+
+error:
+	mtk_vdec_err(inst->ctx, "\n - FRAME[%d] - err=%d\n", inst->frm_cnt, err);
+	return err;
+}
+
+static void get_disp_fb(struct vdec_vp8_inst *inst, struct vdec_fb **out_fb)
+{
+	struct vdec_fb_node *node;
+	struct vdec_fb *fb;
+
+	node = list_first_entry_or_null(&inst->fb_disp_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		list_move_tail(&node->list, &inst->available_fb_node_list);
+		fb = (struct vdec_fb *)node->fb;
+		fb->status |= FB_ST_DISPLAY;
+		mtk_vdec_debug(inst->ctx, "[FB] get disp fb %p st=%d", node->fb, fb->status);
+	} else {
+		fb = NULL;
+		mtk_vdec_debug(inst->ctx, "[FB] there is no disp fb");
+	}
+
+	*out_fb = fb;
+}
+
+static void get_free_fb(struct vdec_vp8_inst *inst, struct vdec_fb **out_fb)
+{
+	struct vdec_fb_node *node;
+	struct vdec_fb *fb;
+
+	node = list_first_entry_or_null(&inst->fb_free_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		list_move_tail(&node->list, &inst->available_fb_node_list);
+		fb = (struct vdec_fb *)node->fb;
+		fb->status |= FB_ST_FREE;
+		mtk_vdec_debug(inst->ctx, "[FB] get free fb %p st=%d", node->fb, fb->status);
+	} else {
+		fb = NULL;
+		mtk_vdec_debug(inst->ctx, "[FB] there is no free fb");
+	}
+
+	*out_fb = fb;
+}
+
+static void get_crop_info(struct vdec_vp8_inst *inst, struct v4l2_rect *cr)
+{
+	cr->left = 0;
+	cr->top = 0;
+	cr->width = inst->vsi->pic.pic_w;
+	cr->height = inst->vsi->pic.pic_h;
+	mtk_vdec_debug(inst->ctx, "get crop info l=%d, t=%d, w=%d, h=%d",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static int vdec_vp8_get_param(void *h_vdec, enum vdec_get_param_type type,
+			      void *out)
+{
+	struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec;
+
+	switch (type) {
+	case GET_PARAM_DISP_FRAME_BUFFER:
+		get_disp_fb(inst, out);
+		break;
+
+	case GET_PARAM_FREE_FRAME_BUFFER:
+		get_free_fb(inst, out);
+		break;
+
+	case GET_PARAM_PIC_INFO:
+		get_pic_info(inst, out);
+		break;
+
+	case GET_PARAM_CROP_INFO:
+		get_crop_info(inst, out);
+		break;
+
+	case GET_PARAM_DPB_SIZE:
+		*((unsigned int *)out) = VP8_DPB_SIZE;
+		break;
+
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void vdec_vp8_deinit(void *h_vdec)
+{
+	struct vdec_vp8_inst *inst = (struct vdec_vp8_inst *)h_vdec;
+
+	vpu_dec_deinit(&inst->vpu);
+	free_working_buf(inst);
+	kfree(inst);
+}
+
+const struct vdec_common_if vdec_vp8_if = {
+	.init		= vdec_vp8_init,
+	.decode		= vdec_vp8_decode,
+	.get_param	= vdec_vp8_get_param,
+	.deinit		= vdec_vp8_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c
new file mode 100644
index 0000000000..f64b21c071
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_req_if.c
@@ -0,0 +1,433 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include <linux/slab.h>
+#include <media/v4l2-mem2mem.h>
+#include <media/videobuf2-dma-contig.h>
+#include <uapi/linux/v4l2-controls.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+
+/* Decoding picture buffer size (3 reference frames plus current frame) */
+#define VP8_DPB_SIZE 4
+
+/* HW working buffer size (bytes) */
+#define VP8_SEG_ID_SZ   SZ_256K
+#define VP8_PP_WRAPY_SZ SZ_64K
+#define VP8_PP_WRAPC_SZ SZ_64K
+#define VP8_VLD_PRED_SZ SZ_64K
+
+/**
+ * struct vdec_vp8_slice_info - decode misc information
+ *
+ * @vld_wrapper_dma:	vld wrapper dma address
+ * @seg_id_buf_dma:	seg id dma address
+ * @wrap_y_dma:	wrap y dma address
+ * @wrap_c_dma:	wrap y dma address
+ * @cur_y_fb_dma:	current plane Y frame buffer dma address
+ * @cur_c_fb_dma:	current plane C frame buffer dma address
+ * @bs_dma:		bitstream dma address
+ * @bs_sz:		bitstream size
+ * @resolution_changed:resolution change flag 1 - changed,  0 - not change
+ * @frame_header_type:	current frame header type
+ * @wait_key_frame:	wait key frame coming
+ * @crc:		used to check whether hardware's status is right
+ * @reserved:		reserved, currently unused
+ */
+struct vdec_vp8_slice_info {
+	u64 vld_wrapper_dma;
+	u64 seg_id_buf_dma;
+	u64 wrap_y_dma;
+	u64 wrap_c_dma;
+	u64 cur_y_fb_dma;
+	u64 cur_c_fb_dma;
+	u64 bs_dma;
+	u32 bs_sz;
+	u32 resolution_changed;
+	u32 frame_header_type;
+	u32 crc[8];
+	u32 reserved;
+};
+
+/**
+ * struct vdec_vp8_slice_dpb_info  - vp8 reference information
+ *
+ * @y_dma_addr:	Y bitstream physical address
+ * @c_dma_addr:	CbCr bitstream physical address
+ * @reference_flag:	reference picture flag
+ * @reserved:		64bit align
+ */
+struct vdec_vp8_slice_dpb_info {
+	dma_addr_t y_dma_addr;
+	dma_addr_t c_dma_addr;
+	int reference_flag;
+	int reserved;
+};
+
+/**
+ * struct vdec_vp8_slice_vsi - VPU shared information
+ *
+ * @dec:		decoding information
+ * @pic:		picture information
+ * @vp8_dpb_info:	reference buffer information
+ */
+struct vdec_vp8_slice_vsi {
+	struct vdec_vp8_slice_info dec;
+	struct vdec_pic_info pic;
+	struct vdec_vp8_slice_dpb_info vp8_dpb_info[3];
+};
+
+/**
+ * struct vdec_vp8_slice_inst - VP8 decoder instance
+ *
+ * @seg_id_buf:	seg buffer
+ * @wrap_y_buf:	wrapper y buffer
+ * @wrap_c_buf:	wrapper c buffer
+ * @vld_wrapper_buf:	vld wrapper buffer
+ * @ctx:		V4L2 context
+ * @vpu:		VPU instance for decoder
+ * @vsi:		VPU share information
+ */
+struct vdec_vp8_slice_inst {
+	struct mtk_vcodec_mem seg_id_buf;
+	struct mtk_vcodec_mem wrap_y_buf;
+	struct mtk_vcodec_mem wrap_c_buf;
+	struct mtk_vcodec_mem vld_wrapper_buf;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vpu_inst vpu;
+	struct vdec_vp8_slice_vsi *vsi;
+};
+
+static void *vdec_vp8_slice_get_ctrl_ptr(struct mtk_vcodec_dec_ctx *ctx, int id)
+{
+	struct v4l2_ctrl *ctrl = v4l2_ctrl_find(&ctx->ctrl_hdl, id);
+
+	if (!ctrl)
+		return ERR_PTR(-EINVAL);
+
+	return ctrl->p_cur.p;
+}
+
+static void vdec_vp8_slice_get_pic_info(struct vdec_vp8_slice_inst *inst)
+{
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+	unsigned int data[3];
+
+	data[0] = ctx->picinfo.pic_w;
+	data[1] = ctx->picinfo.pic_h;
+	data[2] = ctx->capture_fourcc;
+	vpu_dec_get_param(&inst->vpu, data, 3, GET_PARAM_PIC_INFO);
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, 64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, 64);
+	ctx->picinfo.fb_sz[0] = inst->vpu.fb_sz[0];
+	ctx->picinfo.fb_sz[1] = inst->vpu.fb_sz[1];
+
+	inst->vsi->pic.pic_w = ctx->picinfo.pic_w;
+	inst->vsi->pic.pic_h = ctx->picinfo.pic_h;
+	inst->vsi->pic.buf_w = ctx->picinfo.buf_w;
+	inst->vsi->pic.buf_h = ctx->picinfo.buf_h;
+	inst->vsi->pic.fb_sz[0] = ctx->picinfo.fb_sz[0];
+	inst->vsi->pic.fb_sz[1] = ctx->picinfo.fb_sz[1];
+	mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)",
+		       ctx->picinfo.pic_w, ctx->picinfo.pic_h,
+		       ctx->picinfo.buf_w, ctx->picinfo.buf_h);
+	mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)",
+		       ctx->picinfo.fb_sz[0], ctx->picinfo.fb_sz[1]);
+}
+
+static int vdec_vp8_slice_alloc_working_buf(struct vdec_vp8_slice_inst *inst)
+{
+	int err;
+	struct mtk_vcodec_mem *mem;
+
+	mem = &inst->seg_id_buf;
+	mem->size = VP8_SEG_ID_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "Cannot allocate working buffer");
+		return err;
+	}
+	inst->vsi->dec.seg_id_buf_dma = (u64)mem->dma_addr;
+
+	mem = &inst->wrap_y_buf;
+	mem->size = VP8_PP_WRAPY_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "cannot allocate WRAP Y buffer");
+		return err;
+	}
+	inst->vsi->dec.wrap_y_dma = (u64)mem->dma_addr;
+
+	mem = &inst->wrap_c_buf;
+	mem->size = VP8_PP_WRAPC_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "cannot allocate WRAP C buffer");
+		return err;
+	}
+	inst->vsi->dec.wrap_c_dma = (u64)mem->dma_addr;
+
+	mem = &inst->vld_wrapper_buf;
+	mem->size = VP8_VLD_PRED_SZ;
+	err = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (err) {
+		mtk_vdec_err(inst->ctx, "cannot allocate vld wrapper buffer");
+		return err;
+	}
+	inst->vsi->dec.vld_wrapper_dma = (u64)mem->dma_addr;
+
+	return 0;
+}
+
+static void vdec_vp8_slice_free_working_buf(struct vdec_vp8_slice_inst *inst)
+{
+	struct mtk_vcodec_mem *mem;
+
+	mem = &inst->seg_id_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+	inst->vsi->dec.seg_id_buf_dma = 0;
+
+	mem = &inst->wrap_y_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+	inst->vsi->dec.wrap_y_dma = 0;
+
+	mem = &inst->wrap_c_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+	inst->vsi->dec.wrap_c_dma = 0;
+
+	mem = &inst->vld_wrapper_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+	inst->vsi->dec.vld_wrapper_dma = 0;
+}
+
+static u64 vdec_vp8_slice_get_ref_by_ts(const struct v4l2_ctrl_vp8_frame *frame_header,
+					int index)
+{
+	switch (index) {
+	case 0:
+		return frame_header->last_frame_ts;
+	case 1:
+		return frame_header->golden_frame_ts;
+	case 2:
+		return frame_header->alt_frame_ts;
+	default:
+		break;
+	}
+
+	return -1;
+}
+
+static int vdec_vp8_slice_get_decode_parameters(struct vdec_vp8_slice_inst *inst)
+{
+	const struct v4l2_ctrl_vp8_frame *frame_header;
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+	struct vb2_queue *vq;
+	struct vb2_buffer *vb;
+	u64 referenct_ts;
+	int index;
+
+	frame_header = vdec_vp8_slice_get_ctrl_ptr(inst->ctx, V4L2_CID_STATELESS_VP8_FRAME);
+	if (IS_ERR(frame_header))
+		return PTR_ERR(frame_header);
+
+	vq = v4l2_m2m_get_vq(ctx->m2m_ctx, V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	for (index = 0; index < 3; index++) {
+		referenct_ts = vdec_vp8_slice_get_ref_by_ts(frame_header, index);
+		vb = vb2_find_buffer(vq, referenct_ts);
+		if (!vb) {
+			if (!V4L2_VP8_FRAME_IS_KEY_FRAME(frame_header))
+				mtk_vdec_err(inst->ctx, "reference invalid: index(%d) ts(%lld)",
+					     index, referenct_ts);
+			inst->vsi->vp8_dpb_info[index].reference_flag = 0;
+			continue;
+		}
+		inst->vsi->vp8_dpb_info[index].reference_flag = 1;
+
+		inst->vsi->vp8_dpb_info[index].y_dma_addr =
+			vb2_dma_contig_plane_dma_addr(vb, 0);
+		if (ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 2)
+			inst->vsi->vp8_dpb_info[index].c_dma_addr =
+				vb2_dma_contig_plane_dma_addr(vb, 1);
+		else
+			inst->vsi->vp8_dpb_info[index].c_dma_addr =
+				inst->vsi->vp8_dpb_info[index].y_dma_addr +
+				ctx->picinfo.fb_sz[0];
+	}
+
+	inst->vsi->dec.frame_header_type = frame_header->flags >> 1;
+
+	return 0;
+}
+
+static int vdec_vp8_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_vp8_slice_inst *inst;
+	int err;
+
+	inst = kzalloc(sizeof(*inst), GFP_KERNEL);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->ctx = ctx;
+
+	inst->vpu.id = SCP_IPI_VDEC_LAT;
+	inst->vpu.core_id = SCP_IPI_VDEC_CORE;
+	inst->vpu.ctx = ctx;
+	inst->vpu.codec_type = ctx->current_codec;
+	inst->vpu.capture_type = ctx->capture_fourcc;
+
+	err = vpu_dec_init(&inst->vpu);
+	if (err) {
+		mtk_vdec_err(ctx, "vdec_vp8 init err=%d", err);
+		goto error_free_inst;
+	}
+
+	inst->vsi = inst->vpu.vsi;
+	err = vdec_vp8_slice_alloc_working_buf(inst);
+	if (err)
+		goto error_deinit;
+
+	mtk_vdec_debug(ctx, "vp8 struct size = %d vsi: %d\n",
+		       (int)sizeof(struct v4l2_ctrl_vp8_frame),
+		       (int)sizeof(struct vdec_vp8_slice_vsi));
+	mtk_vdec_debug(ctx, "vp8:%p, codec_type = 0x%x vsi: 0x%p",
+		       inst, inst->vpu.codec_type, inst->vpu.vsi);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+error_deinit:
+	vpu_dec_deinit(&inst->vpu);
+error_free_inst:
+	kfree(inst);
+	return err;
+}
+
+static int vdec_vp8_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				 struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp8_slice_inst *inst = h_vdec;
+	struct vdec_vpu_inst *vpu = &inst->vpu;
+	struct mtk_video_dec_buf *src_buf_info, *dst_buf_info;
+	unsigned int data;
+	u64 y_fb_dma, c_fb_dma;
+	int err, timeout;
+
+	/* Resolution changes are never initiated by us */
+	*res_chg = false;
+
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vpu_dec_reset(vpu);
+
+	src_buf_info = container_of(bs, struct mtk_video_dec_buf, bs_buffer);
+
+	fb = inst->ctx->dev->vdec_pdata->get_cap_buffer(inst->ctx);
+	dst_buf_info = container_of(fb, struct mtk_video_dec_buf, frame_buffer);
+
+	y_fb_dma = fb ? (u64)fb->base_y.dma_addr : 0;
+	if (inst->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes == 1)
+		c_fb_dma = y_fb_dma +
+			inst->ctx->picinfo.buf_w * inst->ctx->picinfo.buf_h;
+	else
+		c_fb_dma = fb ? (u64)fb->base_c.dma_addr : 0;
+
+	inst->vsi->dec.bs_dma = (u64)bs->dma_addr;
+	inst->vsi->dec.bs_sz = bs->size;
+	inst->vsi->dec.cur_y_fb_dma = y_fb_dma;
+	inst->vsi->dec.cur_c_fb_dma = c_fb_dma;
+
+	mtk_vdec_debug(inst->ctx, "frame[%d] bs(%zu 0x%llx) y/c(0x%llx 0x%llx)",
+		       inst->ctx->decoded_frame_cnt,
+		       bs->size, (u64)bs->dma_addr,
+		       y_fb_dma, c_fb_dma);
+
+	v4l2_m2m_buf_copy_metadata(&src_buf_info->m2m_buf.vb,
+				   &dst_buf_info->m2m_buf.vb, true);
+
+	err = vdec_vp8_slice_get_decode_parameters(inst);
+	if (err)
+		goto error;
+
+	err = vpu_dec_start(vpu, &data, 1);
+	if (err) {
+		mtk_vdec_debug(inst->ctx, "vp8 dec start err!");
+		goto error;
+	}
+
+	if (inst->vsi->dec.resolution_changed) {
+		mtk_vdec_debug(inst->ctx, "- resolution_changed -");
+		*res_chg = true;
+		return 0;
+	}
+
+	/* wait decode done interrupt */
+	timeout = mtk_vcodec_wait_for_done_ctx(inst->ctx, MTK_INST_IRQ_RECEIVED,
+					       50, MTK_VDEC_CORE);
+
+	err = vpu_dec_end(vpu);
+	if (err || timeout)
+		mtk_vdec_debug(inst->ctx, "vp8 dec error timeout:%d err: %d pic_%d",
+			       timeout, err, inst->ctx->decoded_frame_cnt);
+
+	mtk_vdec_debug(inst->ctx, "pic[%d] crc: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x",
+		       inst->ctx->decoded_frame_cnt,
+		       inst->vsi->dec.crc[0], inst->vsi->dec.crc[1],
+		       inst->vsi->dec.crc[2], inst->vsi->dec.crc[3],
+		       inst->vsi->dec.crc[4], inst->vsi->dec.crc[5],
+		       inst->vsi->dec.crc[6], inst->vsi->dec.crc[7]);
+
+	inst->ctx->decoded_frame_cnt++;
+error:
+	return err;
+}
+
+static int vdec_vp8_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out)
+{
+	struct vdec_vp8_slice_inst *inst = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		vdec_vp8_slice_get_pic_info(inst);
+		break;
+	case GET_PARAM_CROP_INFO:
+		mtk_vdec_debug(inst->ctx, "No need to get vp8 crop information.");
+		break;
+	case GET_PARAM_DPB_SIZE:
+		*((unsigned int *)out) = VP8_DPB_SIZE;
+		break;
+	default:
+		mtk_vdec_err(inst->ctx, "invalid get parameter type=%d", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void vdec_vp8_slice_deinit(void *h_vdec)
+{
+	struct vdec_vp8_slice_inst *inst = h_vdec;
+
+	vpu_dec_deinit(&inst->vpu);
+	vdec_vp8_slice_free_working_buf(inst);
+	kfree(inst);
+}
+
+const struct vdec_common_if vdec_vp8_slice_if = {
+	.init		= vdec_vp8_slice_init,
+	.decode		= vdec_vp8_slice_decode,
+	.get_param	= vdec_vp8_slice_get_param,
+	.deinit		= vdec_vp8_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c
new file mode 100644
index 0000000000..55355fa700
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c
@@ -0,0 +1,1017 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: Daniel Hsiao <daniel.hsiao@mediatek.com>
+ *	Kai-Sean Yang <kai-sean.yang@mediatek.com>
+ *	Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#include <linux/fs.h>
+#include <linux/slab.h>
+#include <linux/syscalls.h>
+#include <linux/delay.h>
+#include <linux/time.h>
+
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_vpu_if.h"
+
+#define VP9_SUPER_FRAME_BS_SZ 64
+#define MAX_VP9_DPB_SIZE	9
+
+#define REFS_PER_FRAME 3
+#define MAX_NUM_REF_FRAMES 8
+#define VP9_MAX_FRM_BUF_NUM 9
+#define VP9_MAX_FRM_BUF_NODE_NUM (VP9_MAX_FRM_BUF_NUM * 2)
+#define VP9_SEG_ID_SZ 0x12000
+
+/**
+ * struct vp9_dram_buf - contains buffer info for vpu
+ * @va : cpu address
+ * @pa : iova address
+ * @sz : buffer size
+ * @padding : for 64 bytes alignment
+ */
+struct vp9_dram_buf {
+	unsigned long va;
+	unsigned long pa;
+	unsigned int sz;
+	unsigned int padding;
+};
+
+/**
+ * struct vp9_fb_info - contains frame buffer info
+ * @fb : frmae buffer
+ * @reserved : reserved field used by vpu
+ */
+struct vp9_fb_info {
+	struct vdec_fb *fb;
+	unsigned int reserved[32];
+};
+
+/**
+ * struct vp9_ref_cnt_buf - contains reference buffer information
+ * @buf : referenced frame buffer
+ * @ref_cnt : referenced frame buffer's reference count.
+ *	When reference count=0, remove it from reference list
+ */
+struct vp9_ref_cnt_buf {
+	struct vp9_fb_info buf;
+	unsigned int ref_cnt;
+};
+
+/**
+ * struct vp9_ref_buf - contains current frame's reference buffer information
+ * @buf : reference buffer
+ * @idx : reference buffer index to frm_bufs
+ * @reserved : reserved field used by vpu
+ */
+struct vp9_ref_buf {
+	struct vp9_fb_info *buf;
+	unsigned int idx;
+	unsigned int reserved[6];
+};
+
+/**
+ * struct vp9_sf_ref_fb - contains frame buffer info
+ * @fb : super frame reference frame buffer
+ * @used : this reference frame info entry is used
+ * @padding : for 64 bytes size align
+ */
+struct vp9_sf_ref_fb {
+	struct vdec_fb fb;
+	int used;
+	int padding;
+};
+
+/*
+ * struct vdec_vp9_vsi - shared buffer between host and VPU firmware
+ *	AP-W/R : AP is writer/reader on this item
+ *	VPU-W/R: VPU is write/reader on this item
+ * @sf_bs_buf : super frame backup buffer (AP-W, VPU-R)
+ * @sf_ref_fb : record supoer frame reference buffer information
+ *	(AP-R/W, VPU-R/W)
+ * @sf_next_ref_fb_idx : next available super frame (AP-W, VPU-R)
+ * @sf_frm_cnt : super frame count, filled by vpu (AP-R, VPU-W)
+ * @sf_frm_offset : super frame offset, filled by vpu (AP-R, VPU-W)
+ * @sf_frm_sz : super frame size, filled by vpu (AP-R, VPU-W)
+ * @sf_frm_idx : current super frame (AP-R, VPU-W)
+ * @sf_init : inform super frame info already parsed by vpu (AP-R, VPU-W)
+ * @fb : capture buffer (AP-W, VPU-R)
+ * @bs : bs buffer (AP-W, VPU-R)
+ * @cur_fb : current show capture buffer (AP-R/W, VPU-R/W)
+ * @pic_w : picture width (AP-R, VPU-W)
+ * @pic_h : picture height (AP-R, VPU-W)
+ * @buf_w : codec width (AP-R, VPU-W)
+ * @buf_h : coded height (AP-R, VPU-W)
+ * @buf_sz_y_bs : ufo compressed y plane size (AP-R, VPU-W)
+ * @buf_sz_c_bs : ufo compressed cbcr plane size (AP-R, VPU-W)
+ * @buf_len_sz_y : size used to store y plane ufo info (AP-R, VPU-W)
+ * @buf_len_sz_c : size used to store cbcr plane ufo info (AP-R, VPU-W)
+
+ * @profile : profile sparsed from vpu (AP-R, VPU-W)
+ * @show_frame : [BIT(0)] display this frame or not (AP-R, VPU-W)
+ *	[BIT(1)] reset segment data or not (AP-R, VPU-W)
+ *	[BIT(2)] trig decoder hardware or not (AP-R, VPU-W)
+ *	[BIT(3)] ask VPU to set bits(0~4) accordingly (AP-W, VPU-R)
+ *	[BIT(4)] do not reset segment data before every frame (AP-R, VPU-W)
+ * @show_existing_frame : inform this frame is show existing frame
+ *	(AP-R, VPU-W)
+ * @frm_to_show_idx : index to show frame (AP-R, VPU-W)
+
+ * @refresh_frm_flags : indicate when frame need to refine reference count
+ *	(AP-R, VPU-W)
+ * @resolution_changed : resolution change in this frame (AP-R, VPU-W)
+
+ * @frm_bufs : maintain reference buffer info (AP-R/W, VPU-R/W)
+ * @ref_frm_map : maintain reference buffer map info (AP-R/W, VPU-R/W)
+ * @new_fb_idx : index to frm_bufs array (AP-R, VPU-W)
+ * @frm_num : decoded frame number, include sub-frame count (AP-R, VPU-W)
+ * @mv_buf : motion vector working buffer (AP-W, VPU-R)
+ * @frm_refs : maintain three reference buffer info (AP-R/W, VPU-R/W)
+ * @seg_id_buf : segmentation map working buffer (AP-W, VPU-R)
+ */
+struct vdec_vp9_vsi {
+	unsigned char sf_bs_buf[VP9_SUPER_FRAME_BS_SZ];
+	struct vp9_sf_ref_fb sf_ref_fb[VP9_MAX_FRM_BUF_NUM-1];
+	int sf_next_ref_fb_idx;
+	unsigned int sf_frm_cnt;
+	unsigned int sf_frm_offset[VP9_MAX_FRM_BUF_NUM-1];
+	unsigned int sf_frm_sz[VP9_MAX_FRM_BUF_NUM-1];
+	unsigned int sf_frm_idx;
+	unsigned int sf_init;
+	struct vdec_fb fb;
+	struct mtk_vcodec_mem bs;
+	struct vdec_fb cur_fb;
+	unsigned int pic_w;
+	unsigned int pic_h;
+	unsigned int buf_w;
+	unsigned int buf_h;
+	unsigned int buf_sz_y_bs;
+	unsigned int buf_sz_c_bs;
+	unsigned int buf_len_sz_y;
+	unsigned int buf_len_sz_c;
+	unsigned int profile;
+	unsigned int show_frame;
+	unsigned int show_existing_frame;
+	unsigned int frm_to_show_idx;
+	unsigned int refresh_frm_flags;
+	unsigned int resolution_changed;
+
+	struct vp9_ref_cnt_buf frm_bufs[VP9_MAX_FRM_BUF_NUM];
+	int ref_frm_map[MAX_NUM_REF_FRAMES];
+	unsigned int new_fb_idx;
+	unsigned int frm_num;
+	struct vp9_dram_buf mv_buf;
+
+	struct vp9_ref_buf frm_refs[REFS_PER_FRAME];
+	struct vp9_dram_buf seg_id_buf;
+
+};
+
+/*
+ * struct vdec_vp9_inst - vp9 decode instance
+ * @mv_buf : working buffer for mv
+ * @seg_id_buf : working buffer for segmentation map
+ * @dec_fb : vdec_fb node to link fb to different fb_xxx_list
+ * @available_fb_node_list : current available vdec_fb node
+ * @fb_use_list : current used or referenced vdec_fb
+ * @fb_free_list : current available to free vdec_fb
+ * @fb_disp_list : current available to display vdec_fb
+ * @cur_fb : current frame buffer
+ * @ctx : current decode context
+ * @vpu : vpu instance information
+ * @vsi : shared buffer between host and VPU firmware
+ * @total_frm_cnt : total frame count, it do not include sub-frames in super
+ *	    frame
+ * @mem : instance memory information
+ */
+struct vdec_vp9_inst {
+	struct mtk_vcodec_mem mv_buf;
+	struct mtk_vcodec_mem seg_id_buf;
+
+	struct vdec_fb_node dec_fb[VP9_MAX_FRM_BUF_NODE_NUM];
+	struct list_head available_fb_node_list;
+	struct list_head fb_use_list;
+	struct list_head fb_free_list;
+	struct list_head fb_disp_list;
+	struct vdec_fb *cur_fb;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vpu_inst vpu;
+	struct vdec_vp9_vsi *vsi;
+	unsigned int total_frm_cnt;
+	struct mtk_vcodec_mem mem;
+};
+
+static bool vp9_is_sf_ref_fb(struct vdec_vp9_inst *inst, struct vdec_fb *fb)
+{
+	int i;
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+
+	for (i = 0; i < ARRAY_SIZE(vsi->sf_ref_fb); i++) {
+		if (fb == &vsi->sf_ref_fb[i].fb)
+			return true;
+	}
+	return false;
+}
+
+static struct vdec_fb *vp9_rm_from_fb_use_list(struct vdec_vp9_inst
+					*inst, void *addr)
+{
+	struct vdec_fb *fb = NULL;
+	struct vdec_fb_node *node;
+
+	list_for_each_entry(node, &inst->fb_use_list, list) {
+		fb = (struct vdec_fb *)node->fb;
+		if (fb->base_y.va == addr) {
+			list_move_tail(&node->list,
+				       &inst->available_fb_node_list);
+			return fb;
+		}
+	}
+
+	return NULL;
+}
+
+static void vp9_add_to_fb_free_list(struct vdec_vp9_inst *inst,
+			     struct vdec_fb *fb)
+{
+	struct vdec_fb_node *node;
+
+	if (fb) {
+		node = list_first_entry_or_null(&inst->available_fb_node_list,
+					struct vdec_fb_node, list);
+
+		if (node) {
+			node->fb = fb;
+			list_move_tail(&node->list, &inst->fb_free_list);
+		}
+	} else {
+		mtk_vdec_debug(inst->ctx, "No free fb node");
+	}
+}
+
+static void vp9_free_sf_ref_fb(struct vdec_fb *fb)
+{
+	struct vp9_sf_ref_fb *sf_ref_fb =
+		container_of(fb, struct vp9_sf_ref_fb, fb);
+
+	sf_ref_fb->used = 0;
+}
+
+static void vp9_ref_cnt_fb(struct vdec_vp9_inst *inst, int *idx,
+			   int new_idx)
+{
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+	int ref_idx = *idx;
+
+	if (ref_idx >= 0 && vsi->frm_bufs[ref_idx].ref_cnt > 0) {
+		vsi->frm_bufs[ref_idx].ref_cnt--;
+
+		if (vsi->frm_bufs[ref_idx].ref_cnt == 0) {
+			if (!vp9_is_sf_ref_fb(inst,
+					      vsi->frm_bufs[ref_idx].buf.fb)) {
+				struct vdec_fb *fb;
+
+				fb = vp9_rm_from_fb_use_list(inst,
+				     vsi->frm_bufs[ref_idx].buf.fb->base_y.va);
+				vp9_add_to_fb_free_list(inst, fb);
+			} else
+				vp9_free_sf_ref_fb(
+					vsi->frm_bufs[ref_idx].buf.fb);
+		}
+	}
+
+	*idx = new_idx;
+	vsi->frm_bufs[new_idx].ref_cnt++;
+}
+
+static void vp9_free_all_sf_ref_fb(struct vdec_vp9_inst *inst)
+{
+	int i;
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+
+	for (i = 0; i < ARRAY_SIZE(vsi->sf_ref_fb); i++) {
+		if (vsi->sf_ref_fb[i].fb.base_y.va) {
+			mtk_vcodec_mem_free(inst->ctx,
+				&vsi->sf_ref_fb[i].fb.base_y);
+			mtk_vcodec_mem_free(inst->ctx,
+				&vsi->sf_ref_fb[i].fb.base_c);
+			vsi->sf_ref_fb[i].used = 0;
+		}
+	}
+}
+
+/* For each sub-frame except the last one, the driver will dynamically
+ * allocate reference buffer by calling vp9_get_sf_ref_fb()
+ * The last sub-frame will use the original fb provided by the
+ * vp9_dec_decode() interface
+ */
+static int vp9_get_sf_ref_fb(struct vdec_vp9_inst *inst)
+{
+	int idx;
+	struct mtk_vcodec_mem *mem_basy_y;
+	struct mtk_vcodec_mem *mem_basy_c;
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+
+	for (idx = 0;
+		idx < ARRAY_SIZE(vsi->sf_ref_fb);
+		idx++) {
+		if (vsi->sf_ref_fb[idx].fb.base_y.va &&
+		    vsi->sf_ref_fb[idx].used == 0) {
+			return idx;
+		}
+	}
+
+	for (idx = 0;
+		idx < ARRAY_SIZE(vsi->sf_ref_fb);
+		idx++) {
+		if (vsi->sf_ref_fb[idx].fb.base_y.va == NULL)
+			break;
+	}
+
+	if (idx == ARRAY_SIZE(vsi->sf_ref_fb)) {
+		mtk_vdec_err(inst->ctx, "List Full");
+		return -1;
+	}
+
+	mem_basy_y = &vsi->sf_ref_fb[idx].fb.base_y;
+	mem_basy_y->size = vsi->buf_sz_y_bs +
+		vsi->buf_len_sz_y;
+
+	if (mtk_vcodec_mem_alloc(inst->ctx, mem_basy_y)) {
+		mtk_vdec_err(inst->ctx, "Cannot allocate sf_ref_buf y_buf");
+		return -1;
+	}
+
+	mem_basy_c = &vsi->sf_ref_fb[idx].fb.base_c;
+	mem_basy_c->size = vsi->buf_sz_c_bs +
+		vsi->buf_len_sz_c;
+
+	if (mtk_vcodec_mem_alloc(inst->ctx, mem_basy_c)) {
+		mtk_vdec_err(inst->ctx, "Cannot allocate sf_ref_fb c_buf");
+		return -1;
+	}
+	vsi->sf_ref_fb[idx].used = 0;
+
+	return idx;
+}
+
+static bool vp9_alloc_work_buf(struct vdec_vp9_inst *inst)
+{
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+	int result;
+	struct mtk_vcodec_mem *mem;
+
+	unsigned int max_pic_w;
+	unsigned int max_pic_h;
+
+
+	if (!(inst->ctx->dev->dec_capability &
+		VCODEC_CAPABILITY_4K_DISABLED)) {
+		max_pic_w = VCODEC_DEC_4K_CODED_WIDTH;
+		max_pic_h = VCODEC_DEC_4K_CODED_HEIGHT;
+	} else {
+		max_pic_w = MTK_VDEC_MAX_W;
+		max_pic_h = MTK_VDEC_MAX_H;
+	}
+
+	if ((vsi->pic_w > max_pic_w) ||
+		(vsi->pic_h > max_pic_h)) {
+		mtk_vdec_err(inst->ctx, "Invalid w/h %d/%d", vsi->pic_w, vsi->pic_h);
+		return false;
+	}
+
+	mtk_vdec_debug(inst->ctx, "BUF CHG(%d): w/h/sb_w/sb_h=%d/%d/%d/%d",
+		       vsi->resolution_changed, vsi->pic_w,
+		       vsi->pic_h, vsi->buf_w, vsi->buf_h);
+
+	mem = &inst->mv_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	mem->size = ((vsi->buf_w / 64) *
+		    (vsi->buf_h / 64) + 2) * 36 * 16;
+	result = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (result) {
+		mem->size = 0;
+		mtk_vdec_err(inst->ctx, "Cannot allocate mv_buf");
+		return false;
+	}
+	/* Set the va again */
+	vsi->mv_buf.va = (unsigned long)mem->va;
+	vsi->mv_buf.pa = (unsigned long)mem->dma_addr;
+	vsi->mv_buf.sz = (unsigned int)mem->size;
+
+
+	mem = &inst->seg_id_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	mem->size = VP9_SEG_ID_SZ;
+	result = mtk_vcodec_mem_alloc(inst->ctx, mem);
+	if (result) {
+		mem->size = 0;
+		mtk_vdec_err(inst->ctx, "Cannot allocate seg_id_buf");
+		return false;
+	}
+	/* Set the va again */
+	vsi->seg_id_buf.va = (unsigned long)mem->va;
+	vsi->seg_id_buf.pa = (unsigned long)mem->dma_addr;
+	vsi->seg_id_buf.sz = (unsigned int)mem->size;
+
+
+	vp9_free_all_sf_ref_fb(inst);
+	vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst);
+
+	return true;
+}
+
+static bool vp9_add_to_fb_disp_list(struct vdec_vp9_inst *inst,
+			     struct vdec_fb *fb)
+{
+	struct vdec_fb_node *node;
+
+	if (!fb) {
+		mtk_vdec_err(inst->ctx, "fb == NULL");
+		return false;
+	}
+
+	node = list_first_entry_or_null(&inst->available_fb_node_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		node->fb = fb;
+		list_move_tail(&node->list, &inst->fb_disp_list);
+	} else {
+		mtk_vdec_err(inst->ctx, "No available fb node");
+		return false;
+	}
+
+	return true;
+}
+
+/* If any buffer updating is signaled it should be done here. */
+static void vp9_swap_frm_bufs(struct vdec_vp9_inst *inst)
+{
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+	struct vp9_fb_info *frm_to_show;
+	int ref_index = 0, mask;
+
+	for (mask = vsi->refresh_frm_flags; mask; mask >>= 1) {
+		if (mask & 1)
+			vp9_ref_cnt_fb(inst, &vsi->ref_frm_map[ref_index],
+				       vsi->new_fb_idx);
+		++ref_index;
+	}
+
+	frm_to_show = &vsi->frm_bufs[vsi->new_fb_idx].buf;
+	vsi->frm_bufs[vsi->new_fb_idx].ref_cnt--;
+
+	if (frm_to_show->fb != inst->cur_fb) {
+		/* This frame is show exist frame and no decode output
+		 * copy frame data from frm_to_show to current CAPTURE
+		 * buffer
+		 */
+		if ((frm_to_show->fb != NULL) &&
+			(inst->cur_fb->base_y.size >=
+			frm_to_show->fb->base_y.size) &&
+			(inst->cur_fb->base_c.size >=
+			frm_to_show->fb->base_c.size)) {
+			memcpy((void *)inst->cur_fb->base_y.va,
+				(void *)frm_to_show->fb->base_y.va,
+				frm_to_show->fb->base_y.size);
+			memcpy((void *)inst->cur_fb->base_c.va,
+				(void *)frm_to_show->fb->base_c.va,
+				frm_to_show->fb->base_c.size);
+		} else {
+			/* After resolution change case, current CAPTURE buffer
+			 * may have less buffer size than frm_to_show buffer
+			 * size
+			 */
+			if (frm_to_show->fb != NULL)
+				mtk_vdec_err(inst->ctx,
+					     "base_y.size=%zu, frm_to_show: base_y.size=%zu",
+					     inst->cur_fb->base_y.size,
+					     frm_to_show->fb->base_y.size);
+		}
+		if (!vp9_is_sf_ref_fb(inst, inst->cur_fb)) {
+			if (vsi->show_frame & BIT(0))
+				vp9_add_to_fb_disp_list(inst, inst->cur_fb);
+		}
+	} else {
+		if (!vp9_is_sf_ref_fb(inst, inst->cur_fb)) {
+			if (vsi->show_frame & BIT(0))
+				vp9_add_to_fb_disp_list(inst, frm_to_show->fb);
+		}
+	}
+
+	/* when ref_cnt ==0, move this fb to fb_free_list. v4l2 driver will
+	 * clean fb_free_list
+	 */
+	if (vsi->frm_bufs[vsi->new_fb_idx].ref_cnt == 0) {
+		if (!vp9_is_sf_ref_fb(
+			inst, vsi->frm_bufs[vsi->new_fb_idx].buf.fb)) {
+			struct vdec_fb *fb;
+
+			fb = vp9_rm_from_fb_use_list(inst,
+			vsi->frm_bufs[vsi->new_fb_idx].buf.fb->base_y.va);
+
+			vp9_add_to_fb_free_list(inst, fb);
+		} else {
+			vp9_free_sf_ref_fb(
+				vsi->frm_bufs[vsi->new_fb_idx].buf.fb);
+		}
+	}
+
+	/* if this super frame and it is not last sub-frame, get next fb for
+	 * sub-frame decode
+	 */
+	if (vsi->sf_frm_cnt > 0 && vsi->sf_frm_idx != vsi->sf_frm_cnt - 1)
+		vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst);
+}
+
+static bool vp9_wait_dec_end(struct vdec_vp9_inst *inst)
+{
+	struct mtk_vcodec_dec_ctx *ctx = inst->ctx;
+
+	mtk_vcodec_wait_for_done_ctx(inst->ctx,
+			MTK_INST_IRQ_RECEIVED,
+			WAIT_INTR_TIMEOUT_MS, 0);
+
+	if (ctx->irq_status & MTK_VDEC_IRQ_STATUS_DEC_SUCCESS)
+		return true;
+	else
+		return false;
+}
+
+static struct vdec_vp9_inst *vp9_alloc_inst(struct mtk_vcodec_dec_ctx *ctx)
+{
+	int result;
+	struct mtk_vcodec_mem mem;
+	struct vdec_vp9_inst *inst;
+
+	memset(&mem, 0, sizeof(mem));
+	mem.size = sizeof(struct vdec_vp9_inst);
+	result = mtk_vcodec_mem_alloc(ctx, &mem);
+	if (result)
+		return NULL;
+
+	inst = mem.va;
+	inst->mem = mem;
+
+	return inst;
+}
+
+static void vp9_free_inst(struct vdec_vp9_inst *inst)
+{
+	struct mtk_vcodec_mem mem;
+
+	mem = inst->mem;
+	if (mem.va)
+		mtk_vcodec_mem_free(inst->ctx, &mem);
+}
+
+static bool vp9_decode_end_proc(struct vdec_vp9_inst *inst)
+{
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+	bool ret = false;
+
+	if (!vsi->show_existing_frame) {
+		ret = vp9_wait_dec_end(inst);
+		if (!ret) {
+			mtk_vdec_err(inst->ctx, "Decode failed, Decode Timeout @[%d]",
+				     vsi->frm_num);
+			return false;
+		}
+
+		if (vpu_dec_end(&inst->vpu)) {
+			mtk_vdec_err(inst->ctx, "vp9_dec_vpu_end failed");
+			return false;
+		}
+		mtk_vdec_debug(inst->ctx, "Decode Ok @%d (%d/%d)", vsi->frm_num,
+			       vsi->pic_w, vsi->pic_h);
+	} else {
+		mtk_vdec_debug(inst->ctx, "Decode Ok @%d (show_existing_frame)", vsi->frm_num);
+	}
+
+	vp9_swap_frm_bufs(inst);
+	vsi->frm_num++;
+	return true;
+}
+
+static bool vp9_is_last_sub_frm(struct vdec_vp9_inst *inst)
+{
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+
+	if (vsi->sf_frm_cnt <= 0 || vsi->sf_frm_idx == vsi->sf_frm_cnt)
+		return true;
+
+	return false;
+}
+
+static struct vdec_fb *vp9_rm_from_fb_disp_list(struct vdec_vp9_inst *inst)
+{
+	struct vdec_fb_node *node;
+	struct vdec_fb *fb = NULL;
+
+	node = list_first_entry_or_null(&inst->fb_disp_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		fb = (struct vdec_fb *)node->fb;
+		fb->status |= FB_ST_DISPLAY;
+		list_move_tail(&node->list, &inst->available_fb_node_list);
+		mtk_vdec_debug(inst->ctx, "[FB] get disp fb %p st=%d", node->fb, fb->status);
+	} else
+		mtk_vdec_debug(inst->ctx, "[FB] there is no disp fb");
+
+	return fb;
+}
+
+static bool vp9_add_to_fb_use_list(struct vdec_vp9_inst *inst,
+			    struct vdec_fb *fb)
+{
+	struct vdec_fb_node *node;
+
+	if (!fb) {
+		mtk_vdec_debug(inst->ctx, "fb == NULL");
+		return false;
+	}
+
+	node = list_first_entry_or_null(&inst->available_fb_node_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		node->fb = fb;
+		list_move_tail(&node->list, &inst->fb_use_list);
+	} else {
+		mtk_vdec_err(inst->ctx, "No free fb node");
+		return false;
+	}
+	return true;
+}
+
+static void vp9_reset(struct vdec_vp9_inst *inst)
+{
+	struct vdec_fb_node *node, *tmp;
+
+	list_for_each_entry_safe(node, tmp, &inst->fb_use_list, list)
+		list_move_tail(&node->list, &inst->fb_free_list);
+
+	vp9_free_all_sf_ref_fb(inst);
+	inst->vsi->sf_next_ref_fb_idx = vp9_get_sf_ref_fb(inst);
+
+	if (vpu_dec_reset(&inst->vpu))
+		mtk_vdec_err(inst->ctx, "vp9_dec_vpu_reset failed");
+
+	/* Set the va again, since vpu_dec_reset will clear mv_buf in vpu */
+	inst->vsi->mv_buf.va = (unsigned long)inst->mv_buf.va;
+	inst->vsi->mv_buf.pa = (unsigned long)inst->mv_buf.dma_addr;
+	inst->vsi->mv_buf.sz = (unsigned long)inst->mv_buf.size;
+
+	/* Set the va again, since vpu_dec_reset will clear seg_id_buf in vpu */
+	inst->vsi->seg_id_buf.va = (unsigned long)inst->seg_id_buf.va;
+	inst->vsi->seg_id_buf.pa = (unsigned long)inst->seg_id_buf.dma_addr;
+	inst->vsi->seg_id_buf.sz = (unsigned long)inst->seg_id_buf.size;
+
+}
+
+static void init_all_fb_lists(struct vdec_vp9_inst *inst)
+{
+	int i;
+
+	INIT_LIST_HEAD(&inst->available_fb_node_list);
+	INIT_LIST_HEAD(&inst->fb_use_list);
+	INIT_LIST_HEAD(&inst->fb_free_list);
+	INIT_LIST_HEAD(&inst->fb_disp_list);
+
+	for (i = 0; i < ARRAY_SIZE(inst->dec_fb); i++) {
+		INIT_LIST_HEAD(&inst->dec_fb[i].list);
+		inst->dec_fb[i].fb = NULL;
+		list_add_tail(&inst->dec_fb[i].list,
+			      &inst->available_fb_node_list);
+	}
+}
+
+static void get_pic_info(struct vdec_vp9_inst *inst, struct vdec_pic_info *pic)
+{
+	pic->fb_sz[0] = inst->vsi->buf_sz_y_bs + inst->vsi->buf_len_sz_y;
+	pic->fb_sz[1] = inst->vsi->buf_sz_c_bs + inst->vsi->buf_len_sz_c;
+
+	pic->pic_w = inst->vsi->pic_w;
+	pic->pic_h = inst->vsi->pic_h;
+	pic->buf_w = inst->vsi->buf_w;
+	pic->buf_h = inst->vsi->buf_h;
+
+	mtk_vdec_debug(inst->ctx, "pic(%d, %d), buf(%d, %d)",
+		       pic->pic_w, pic->pic_h, pic->buf_w, pic->buf_h);
+	mtk_vdec_debug(inst->ctx, "fb size: Y(%d), C(%d)", pic->fb_sz[0], pic->fb_sz[1]);
+}
+
+static void get_disp_fb(struct vdec_vp9_inst *inst, struct vdec_fb **out_fb)
+{
+
+	*out_fb = vp9_rm_from_fb_disp_list(inst);
+	if (*out_fb)
+		(*out_fb)->status |= FB_ST_DISPLAY;
+}
+
+static void get_free_fb(struct vdec_vp9_inst *inst, struct vdec_fb **out_fb)
+{
+	struct vdec_fb_node *node;
+	struct vdec_fb *fb = NULL;
+
+	node = list_first_entry_or_null(&inst->fb_free_list,
+					struct vdec_fb_node, list);
+	if (node) {
+		list_move_tail(&node->list, &inst->available_fb_node_list);
+		fb = (struct vdec_fb *)node->fb;
+		fb->status |= FB_ST_FREE;
+		mtk_vdec_debug(inst->ctx, "[FB] get free fb %p st=%d", node->fb, fb->status);
+	} else {
+		mtk_vdec_debug(inst->ctx, "[FB] there is no free fb");
+	}
+
+	*out_fb = fb;
+}
+
+static int validate_vsi_array_indexes(struct vdec_vp9_inst *inst,
+		struct vdec_vp9_vsi *vsi) {
+	if (vsi->sf_frm_idx >= VP9_MAX_FRM_BUF_NUM - 1) {
+		mtk_vdec_err(inst->ctx, "Invalid vsi->sf_frm_idx=%u.", vsi->sf_frm_idx);
+		return -EIO;
+	}
+	if (vsi->frm_to_show_idx >= VP9_MAX_FRM_BUF_NUM) {
+		mtk_vdec_err(inst->ctx, "Invalid vsi->frm_to_show_idx=%u.", vsi->frm_to_show_idx);
+		return -EIO;
+	}
+	if (vsi->new_fb_idx >= VP9_MAX_FRM_BUF_NUM) {
+		mtk_vdec_err(inst->ctx, "Invalid vsi->new_fb_idx=%u.", vsi->new_fb_idx);
+		return -EIO;
+	}
+	return 0;
+}
+
+static void vdec_vp9_deinit(void *h_vdec)
+{
+	struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec;
+	struct mtk_vcodec_mem *mem;
+	int ret = 0;
+
+	ret = vpu_dec_deinit(&inst->vpu);
+	if (ret)
+		mtk_vdec_err(inst->ctx, "vpu_dec_deinit failed");
+
+	mem = &inst->mv_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	mem = &inst->seg_id_buf;
+	if (mem->va)
+		mtk_vcodec_mem_free(inst->ctx, mem);
+
+	vp9_free_all_sf_ref_fb(inst);
+	vp9_free_inst(inst);
+}
+
+static int vdec_vp9_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_vp9_inst *inst;
+
+	inst = vp9_alloc_inst(ctx);
+	if (!inst)
+		return -ENOMEM;
+
+	inst->total_frm_cnt = 0;
+	inst->ctx = ctx;
+
+	inst->vpu.id = IPI_VDEC_VP9;
+	inst->vpu.ctx = ctx;
+
+	if (vpu_dec_init(&inst->vpu)) {
+		mtk_vdec_err(inst->ctx, "vp9_dec_vpu_init failed");
+		goto err_deinit_inst;
+	}
+
+	inst->vsi = (struct vdec_vp9_vsi *)inst->vpu.vsi;
+
+	inst->vsi->show_frame |= BIT(3);
+
+	init_all_fb_lists(inst);
+
+	ctx->drv_handle = inst;
+	return 0;
+
+err_deinit_inst:
+	vp9_free_inst(inst);
+
+	return -EINVAL;
+}
+
+static int vdec_vp9_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+			   struct vdec_fb *fb, bool *res_chg)
+{
+	int ret = 0;
+	struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec;
+	struct vdec_vp9_vsi *vsi = inst->vsi;
+	u32 data[3];
+	int i;
+
+	*res_chg = false;
+
+	if ((bs == NULL) && (fb == NULL)) {
+		mtk_vdec_debug(inst->ctx, "[EOS]");
+		vp9_reset(inst);
+		return ret;
+	}
+
+	if (bs == NULL) {
+		mtk_vdec_err(inst->ctx, "bs == NULL");
+		return -EINVAL;
+	}
+
+	mtk_vdec_debug(inst->ctx, "Input BS Size = %zu", bs->size);
+
+	while (1) {
+		struct vdec_fb *cur_fb = NULL;
+
+		data[0] = *((unsigned int *)bs->va);
+		data[1] = *((unsigned int *)(bs->va + 4));
+		data[2] = *((unsigned int *)(bs->va + 8));
+
+		vsi->bs = *bs;
+
+		if (fb)
+			vsi->fb = *fb;
+
+		if (!vsi->sf_init) {
+			unsigned int sf_bs_sz;
+			unsigned int sf_bs_off;
+			unsigned char *sf_bs_src;
+			unsigned char *sf_bs_dst;
+
+			sf_bs_sz = bs->size > VP9_SUPER_FRAME_BS_SZ ?
+					VP9_SUPER_FRAME_BS_SZ : bs->size;
+			sf_bs_off = VP9_SUPER_FRAME_BS_SZ - sf_bs_sz;
+			sf_bs_src = bs->va + bs->size - sf_bs_sz;
+			sf_bs_dst = vsi->sf_bs_buf + sf_bs_off;
+			memcpy(sf_bs_dst, sf_bs_src, sf_bs_sz);
+		} else {
+			if ((vsi->sf_frm_cnt > 0) &&
+				(vsi->sf_frm_idx < vsi->sf_frm_cnt)) {
+				unsigned int idx = vsi->sf_frm_idx;
+
+				memcpy((void *)bs->va,
+					(void *)(bs->va +
+					vsi->sf_frm_offset[idx]),
+					vsi->sf_frm_sz[idx]);
+			}
+		}
+
+		if (!(vsi->show_frame & BIT(4)))
+			memset(inst->seg_id_buf.va, 0, inst->seg_id_buf.size);
+
+		ret = vpu_dec_start(&inst->vpu, data, 3);
+		if (ret) {
+			mtk_vdec_err(inst->ctx, "vpu_dec_start failed");
+			goto DECODE_ERROR;
+		}
+
+		if (vsi->show_frame & BIT(1)) {
+			memset(inst->seg_id_buf.va, 0, inst->seg_id_buf.size);
+
+			if (vsi->show_frame & BIT(2)) {
+				ret = vpu_dec_start(&inst->vpu, NULL, 0);
+				if (ret) {
+					mtk_vdec_err(inst->ctx, "vpu trig decoder failed");
+					goto DECODE_ERROR;
+				}
+			}
+		}
+
+		ret = validate_vsi_array_indexes(inst, vsi);
+		if (ret) {
+			mtk_vdec_err(inst->ctx, "Invalid values from VPU.");
+			goto DECODE_ERROR;
+		}
+
+		if (vsi->resolution_changed) {
+			if (!vp9_alloc_work_buf(inst)) {
+				ret = -EIO;
+				goto DECODE_ERROR;
+			}
+		}
+
+		if (vsi->sf_frm_cnt > 0) {
+			cur_fb = &vsi->sf_ref_fb[vsi->sf_next_ref_fb_idx].fb;
+
+			if (vsi->sf_frm_idx < vsi->sf_frm_cnt)
+				inst->cur_fb = cur_fb;
+			else
+				inst->cur_fb = fb;
+		} else {
+			inst->cur_fb = fb;
+		}
+
+		vsi->frm_bufs[vsi->new_fb_idx].buf.fb = inst->cur_fb;
+		if (!vp9_is_sf_ref_fb(inst, inst->cur_fb))
+			vp9_add_to_fb_use_list(inst, inst->cur_fb);
+
+		mtk_vdec_debug(inst->ctx, "[#pic %d]", vsi->frm_num);
+
+		if (vsi->show_existing_frame)
+			mtk_vdec_debug(inst->ctx,
+				       "drv->new_fb_idx=%d, drv->frm_to_show_idx=%d",
+				       vsi->new_fb_idx, vsi->frm_to_show_idx);
+
+		if (vsi->show_existing_frame && (vsi->frm_to_show_idx <
+					VP9_MAX_FRM_BUF_NUM)) {
+			mtk_vdec_debug(inst->ctx,
+				       "Skip Decode drv->new_fb_idx=%d, drv->frm_to_show_idx=%d",
+				       vsi->new_fb_idx, vsi->frm_to_show_idx);
+
+			vp9_ref_cnt_fb(inst, &vsi->new_fb_idx,
+					vsi->frm_to_show_idx);
+		}
+
+		/* VPU assign the buffer pointer in its address space,
+		 * reassign here
+		 */
+		for (i = 0; i < ARRAY_SIZE(vsi->frm_refs); i++) {
+			unsigned int idx = vsi->frm_refs[i].idx;
+
+			vsi->frm_refs[i].buf = &vsi->frm_bufs[idx].buf;
+		}
+
+		if (vsi->resolution_changed) {
+			*res_chg = true;
+			mtk_vdec_debug(inst->ctx, "VDEC_ST_RESOLUTION_CHANGED");
+
+			ret = 0;
+			goto DECODE_ERROR;
+		}
+
+		if (!vp9_decode_end_proc(inst)) {
+			mtk_vdec_err(inst->ctx, "vp9_decode_end_proc");
+			ret = -EINVAL;
+			goto DECODE_ERROR;
+		}
+
+		if (vp9_is_last_sub_frm(inst))
+			break;
+
+	}
+	inst->total_frm_cnt++;
+
+DECODE_ERROR:
+	if (ret < 0)
+		vp9_add_to_fb_free_list(inst, fb);
+
+	return ret;
+}
+
+static void get_crop_info(struct vdec_vp9_inst *inst, struct v4l2_rect *cr)
+{
+	cr->left = 0;
+	cr->top = 0;
+	cr->width = inst->vsi->pic_w;
+	cr->height = inst->vsi->pic_h;
+	mtk_vdec_debug(inst->ctx, "get crop info l=%d, t=%d, w=%d, h=%d\n",
+		       cr->left, cr->top, cr->width, cr->height);
+}
+
+static int vdec_vp9_get_param(void *h_vdec, enum vdec_get_param_type type,
+			      void *out)
+{
+	struct vdec_vp9_inst *inst = (struct vdec_vp9_inst *)h_vdec;
+	int ret = 0;
+
+	switch (type) {
+	case GET_PARAM_DISP_FRAME_BUFFER:
+		get_disp_fb(inst, out);
+		break;
+	case GET_PARAM_FREE_FRAME_BUFFER:
+		get_free_fb(inst, out);
+		break;
+	case GET_PARAM_PIC_INFO:
+		get_pic_info(inst, out);
+		break;
+	case GET_PARAM_DPB_SIZE:
+		*((unsigned int *)out) = MAX_VP9_DPB_SIZE;
+		break;
+	case GET_PARAM_CROP_INFO:
+		get_crop_info(inst, out);
+		break;
+	default:
+		mtk_vdec_err(inst->ctx, "not supported param type %d", type);
+		ret = -EINVAL;
+		break;
+	}
+
+	return ret;
+}
+
+const struct vdec_common_if vdec_vp9_if = {
+	.init		= vdec_vp9_init,
+	.decode		= vdec_vp9_decode,
+	.get_param	= vdec_vp9_get_param,
+	.deinit		= vdec_vp9_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c
new file mode 100644
index 0000000000..e393e3e668
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c
@@ -0,0 +1,2216 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: George Sun <george.sun@mediatek.com>
+ */
+
+#include <linux/module.h>
+#include <linux/slab.h>
+#include <media/videobuf2-dma-contig.h>
+#include <media/v4l2-vp9.h>
+
+#include "../mtk_vcodec_dec.h"
+#include "../../common/mtk_vcodec_intr.h"
+#include "../vdec_drv_base.h"
+#include "../vdec_drv_if.h"
+#include "../vdec_vpu_if.h"
+
+/* reset_frame_context defined in VP9 spec */
+#define VP9_RESET_FRAME_CONTEXT_NONE0 0
+#define VP9_RESET_FRAME_CONTEXT_NONE1 1
+#define VP9_RESET_FRAME_CONTEXT_SPEC 2
+#define VP9_RESET_FRAME_CONTEXT_ALL 3
+
+#define VP9_TILE_BUF_SIZE 4096
+#define VP9_PROB_BUF_SIZE 2560
+#define VP9_COUNTS_BUF_SIZE 16384
+
+#define HDR_FLAG(x) (!!((hdr)->flags & V4L2_VP9_FRAME_FLAG_##x))
+#define LF_FLAG(x) (!!((lf)->flags & V4L2_VP9_LOOP_FILTER_FLAG_##x))
+#define SEG_FLAG(x) (!!((seg)->flags & V4L2_VP9_SEGMENTATION_FLAG_##x))
+#define VP9_BAND_6(band) ((band) == 0 ? 3 : 6)
+
+/*
+ * struct vdec_vp9_slice_frame_ctx - vp9 prob tables footprint
+ */
+struct vdec_vp9_slice_frame_ctx {
+	struct {
+		u8 probs[6][3];
+		u8 padding[2];
+	} coef_probs[4][2][2][6];
+
+	u8 y_mode_prob[4][16];
+	u8 switch_interp_prob[4][16];
+	u8 seg[32];  /* ignore */
+	u8 comp_inter_prob[16];
+	u8 comp_ref_prob[16];
+	u8 single_ref_prob[5][2];
+	u8 single_ref_prob_padding[6];
+
+	u8 joint[3];
+	u8 joint_padding[13];
+	struct {
+		u8 sign;
+		u8 classes[10];
+		u8 padding[5];
+	} sign_classes[2];
+	struct {
+		u8 class0[1];
+		u8 bits[10];
+		u8 padding[5];
+	} class0_bits[2];
+	struct {
+		u8 class0_fp[2][3];
+		u8 fp[3];
+		u8 class0_hp;
+		u8 hp;
+		u8 padding[5];
+	} class0_fp_hp[2];
+
+	u8 uv_mode_prob[10][16];
+	u8 uv_mode_prob_padding[2][16];
+
+	u8 partition_prob[16][4];
+
+	u8 inter_mode_probs[7][4];
+	u8 skip_probs[4];
+
+	u8 tx_p8x8[2][4];
+	u8 tx_p16x16[2][4];
+	u8 tx_p32x32[2][4];
+	u8 intra_inter_prob[8];
+};
+
+/*
+ * struct vdec_vp9_slice_frame_counts - vp9 counts tables footprint
+ */
+struct vdec_vp9_slice_frame_counts {
+	union {
+		struct {
+			u32 band_0[3];
+			u32 padding0[1];
+			u32 band_1_5[5][6];
+			u32 padding1[2];
+		} eob_branch[4][2][2];
+		u32 eob_branch_space[256 * 4];
+	};
+
+	struct {
+		u32 band_0[3][4];
+		u32 band_1_5[5][6][4];
+	} coef_probs[4][2][2];
+
+	u32 intra_inter[4][2];
+	u32 comp_inter[5][2];
+	u32 comp_inter_padding[2];
+	u32 comp_ref[5][2];
+	u32 comp_ref_padding[2];
+	u32 single_ref[5][2][2];
+	u32 inter_mode[7][4];
+	u32 y_mode[4][12];
+	u32 uv_mode[10][10];
+	u32 partition[16][4];
+	u32 switchable_interp[4][4];
+
+	u32 tx_p8x8[2][2];
+	u32 tx_p16x16[2][4];
+	u32 tx_p32x32[2][4];
+
+	u32 skip[3][4];
+
+	u32 joint[4];
+
+	struct {
+		u32 sign[2];
+		u32 class0[2];
+		u32 classes[12];
+		u32 bits[10][2];
+		u32 padding[4];
+		u32 class0_fp[2][4];
+		u32 fp[4];
+		u32 class0_hp[2];
+		u32 hp[2];
+	} mvcomp[2];
+
+	u32 reserved[126][4];
+};
+
+/**
+ * struct vdec_vp9_slice_counts_map - vp9 counts tables to map
+ *                                    v4l2_vp9_frame_symbol_counts
+ * @skip:	skip counts.
+ * @y_mode:	Y prediction mode counts.
+ * @filter:	interpolation filter counts.
+ * @mv_joint:	motion vector joint counts.
+ * @sign:	motion vector sign counts.
+ * @classes:	motion vector class counts.
+ * @class0:	motion vector class0 bit counts.
+ * @bits:	motion vector bits counts.
+ * @class0_fp:	motion vector class0 fractional bit counts.
+ * @fp:	motion vector fractional bit counts.
+ * @class0_hp:	motion vector class0 high precision fractional bit counts.
+ * @hp:	motion vector high precision fractional bit counts.
+ */
+struct vdec_vp9_slice_counts_map {
+	u32 skip[3][2];
+	u32 y_mode[4][10];
+	u32 filter[4][3];
+	u32 sign[2][2];
+	u32 classes[2][11];
+	u32 class0[2][2];
+	u32 bits[2][10][2];
+	u32 class0_fp[2][2][4];
+	u32 fp[2][4];
+	u32 class0_hp[2][2];
+	u32 hp[2][2];
+};
+
+/*
+ * struct vdec_vp9_slice_uncompressed_header - vp9 uncompressed header syntax
+ *                                             used for decoding
+ */
+struct vdec_vp9_slice_uncompressed_header {
+	u8 profile;
+	u8 last_frame_type;
+	u8 frame_type;
+
+	u8 last_show_frame;
+	u8 show_frame;
+	u8 error_resilient_mode;
+
+	u8 bit_depth;
+	u8 padding0[1];
+	u16 last_frame_width;
+	u16 last_frame_height;
+	u16 frame_width;
+	u16 frame_height;
+
+	u8 intra_only;
+	u8 reset_frame_context;
+	u8 ref_frame_sign_bias[4];
+	u8 allow_high_precision_mv;
+	u8 interpolation_filter;
+
+	u8 refresh_frame_context;
+	u8 frame_parallel_decoding_mode;
+	u8 frame_context_idx;
+
+	/* loop_filter_params */
+	u8 loop_filter_level;
+	u8 loop_filter_sharpness;
+	u8 loop_filter_delta_enabled;
+	s8 loop_filter_ref_deltas[4];
+	s8 loop_filter_mode_deltas[2];
+
+	/* quantization_params */
+	u8 base_q_idx;
+	s8 delta_q_y_dc;
+	s8 delta_q_uv_dc;
+	s8 delta_q_uv_ac;
+
+	/* segmentation_params */
+	u8 segmentation_enabled;
+	u8 segmentation_update_map;
+	u8 segmentation_tree_probs[7];
+	u8 padding1[1];
+	u8 segmentation_temporal_udpate;
+	u8 segmentation_pred_prob[3];
+	u8 segmentation_update_data;
+	u8 segmentation_abs_or_delta_update;
+	u8 feature_enabled[8];
+	s16 feature_value[8][4];
+
+	/* tile_info */
+	u8 tile_cols_log2;
+	u8 tile_rows_log2;
+	u8 padding2[2];
+
+	u16 uncompressed_header_size;
+	u16 header_size_in_bytes;
+
+	/* LAT OUT, CORE IN */
+	u32 dequant[8][4];
+};
+
+/*
+ * struct vdec_vp9_slice_compressed_header - vp9 compressed header syntax
+ *                                           used for decoding.
+ */
+struct vdec_vp9_slice_compressed_header {
+	u8 tx_mode;
+	u8 ref_mode;
+	u8 comp_fixed_ref;
+	u8 comp_var_ref[2];
+	u8 padding[3];
+};
+
+/*
+ * struct vdec_vp9_slice_tiles - vp9 tile syntax
+ */
+struct vdec_vp9_slice_tiles {
+	u32 size[4][64];
+	u32 mi_rows[4];
+	u32 mi_cols[64];
+	u8 actual_rows;
+	u8 padding[7];
+};
+
+/*
+ * struct vdec_vp9_slice_reference - vp9 reference frame information
+ */
+struct vdec_vp9_slice_reference {
+	u16 frame_width;
+	u16 frame_height;
+	u8 bit_depth;
+	u8 subsampling_x;
+	u8 subsampling_y;
+	u8 padding;
+};
+
+/*
+ * struct vdec_vp9_slice_frame - vp9 syntax used for decoding
+ */
+struct vdec_vp9_slice_frame {
+	struct vdec_vp9_slice_uncompressed_header uh;
+	struct vdec_vp9_slice_compressed_header ch;
+	struct vdec_vp9_slice_tiles tiles;
+	struct vdec_vp9_slice_reference ref[3];
+};
+
+/*
+ * struct vdec_vp9_slice_init_vsi - VSI used to initialize instance
+ */
+struct vdec_vp9_slice_init_vsi {
+	unsigned int architecture;
+	unsigned int reserved;
+	u64 core_vsi;
+	/* default frame context's position in MicroP */
+	u64 default_frame_ctx;
+};
+
+/*
+ * struct vdec_vp9_slice_mem - memory address and size
+ */
+struct vdec_vp9_slice_mem {
+	union {
+		u64 buf;
+		dma_addr_t dma_addr;
+	};
+	union {
+		size_t size;
+		dma_addr_t dma_addr_end;
+		u64 padding;
+	};
+};
+
+/*
+ * struct vdec_vp9_slice_bs - input buffer for decoding
+ */
+struct vdec_vp9_slice_bs {
+	struct vdec_vp9_slice_mem buf;
+	struct vdec_vp9_slice_mem frame;
+};
+
+/*
+ * struct vdec_vp9_slice_fb - frame buffer for decoding
+ */
+struct vdec_vp9_slice_fb {
+	struct vdec_vp9_slice_mem y;
+	struct vdec_vp9_slice_mem c;
+};
+
+/*
+ * struct vdec_vp9_slice_state - decoding state
+ */
+struct vdec_vp9_slice_state {
+	int err;
+	unsigned int full;
+	unsigned int timeout;
+	unsigned int perf;
+
+	unsigned int crc[12];
+};
+
+/**
+ * struct vdec_vp9_slice_vsi - exchange decoding information
+ *                             between Main CPU and MicroP
+ *
+ * @bs:	input buffer
+ * @fb:	output buffer
+ * @ref:	3 reference buffers
+ * @mv:	mv working buffer
+ * @seg:	segmentation working buffer
+ * @tile:	tile buffer
+ * @prob:	prob table buffer, used to set/update prob table
+ * @counts:	counts table buffer, used to update prob table
+ * @ube:	general buffer
+ * @trans:	trans buffer position in general buffer
+ * @err_map:	error buffer
+ * @row_info:	row info buffer
+ * @frame:	decoding syntax
+ * @state:	decoding state
+ */
+struct vdec_vp9_slice_vsi {
+	/* used in LAT stage */
+	struct vdec_vp9_slice_bs bs;
+	/* used in Core stage */
+	struct vdec_vp9_slice_fb fb;
+	struct vdec_vp9_slice_fb ref[3];
+
+	struct vdec_vp9_slice_mem mv[2];
+	struct vdec_vp9_slice_mem seg[2];
+	struct vdec_vp9_slice_mem tile;
+	struct vdec_vp9_slice_mem prob;
+	struct vdec_vp9_slice_mem counts;
+
+	/* LAT stage's output, Core stage's input */
+	struct vdec_vp9_slice_mem ube;
+	struct vdec_vp9_slice_mem trans;
+	struct vdec_vp9_slice_mem err_map;
+	struct vdec_vp9_slice_mem row_info;
+
+	/* decoding parameters */
+	struct vdec_vp9_slice_frame frame;
+
+	struct vdec_vp9_slice_state state;
+};
+
+/**
+ * struct vdec_vp9_slice_pfc - per-frame context that contains a local vsi.
+ *                             pass it from lat to core
+ *
+ * @vsi:	local vsi. copy to/from remote vsi before/after decoding
+ * @ref_idx:	reference buffer index
+ * @seq:	picture sequence
+ * @state:	decoding state
+ */
+struct vdec_vp9_slice_pfc {
+	struct vdec_vp9_slice_vsi vsi;
+
+	u64 ref_idx[3];
+
+	int seq;
+
+	/* LAT/Core CRC */
+	struct vdec_vp9_slice_state state[2];
+};
+
+/*
+ * enum vdec_vp9_slice_resolution_level
+ */
+enum vdec_vp9_slice_resolution_level {
+	VP9_RES_NONE,
+	VP9_RES_FHD,
+	VP9_RES_4K,
+	VP9_RES_8K,
+};
+
+/*
+ * struct vdec_vp9_slice_ref - picture's width & height should kept
+ *                             for later decoding as reference picture
+ */
+struct vdec_vp9_slice_ref {
+	unsigned int width;
+	unsigned int height;
+};
+
+/**
+ * struct vdec_vp9_slice_instance - represent one vp9 instance
+ *
+ * @ctx:		pointer to codec's context
+ * @vpu:		VPU instance
+ * @seq:		global picture sequence
+ * @level:		level of current resolution
+ * @width:		width of last picture
+ * @height:		height of last picture
+ * @frame_type:	frame_type of last picture
+ * @irq:		irq to Main CPU or MicroP
+ * @show_frame:	show_frame of last picture
+ * @dpb:		picture information (width/height) for reference
+ * @mv:		mv working buffer
+ * @seg:		segmentation working buffer
+ * @tile:		tile buffer
+ * @prob:		prob table buffer, used to set/update prob table
+ * @counts:		counts table buffer, used to update prob table
+ * @frame_ctx:		4 frame context according to VP9 Spec
+ * @frame_ctx_helper:	4 frame context according to newest kernel spec
+ * @dirty:		state of each frame context
+ * @init_vsi:		vsi used for initialized VP9 instance
+ * @vsi:		vsi used for decoding/flush ...
+ * @core_vsi:		vsi used for Core stage
+ *
+ * @sc_pfc:		per frame context single core
+ * @counts_map:	used map to counts_helper
+ * @counts_helper:	counts table according to newest kernel spec
+ */
+struct vdec_vp9_slice_instance {
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vpu_inst vpu;
+
+	int seq;
+
+	enum vdec_vp9_slice_resolution_level level;
+
+	/* for resolution change and get_pic_info */
+	unsigned int width;
+	unsigned int height;
+
+	/* for last_frame_type */
+	unsigned int frame_type;
+	unsigned int irq;
+
+	unsigned int show_frame;
+
+	/* maintain vp9 reference frame state */
+	struct vdec_vp9_slice_ref dpb[VB2_MAX_FRAME];
+
+	/*
+	 * normal working buffers
+	 * mv[0]/seg[0]/tile/prob/counts is used for LAT
+	 * mv[1]/seg[1] is used for CORE
+	 */
+	struct mtk_vcodec_mem mv[2];
+	struct mtk_vcodec_mem seg[2];
+	struct mtk_vcodec_mem tile;
+	struct mtk_vcodec_mem prob;
+	struct mtk_vcodec_mem counts;
+
+	/* 4 prob tables */
+	struct vdec_vp9_slice_frame_ctx frame_ctx[4];
+	/*4 helper tables */
+	struct v4l2_vp9_frame_context frame_ctx_helper;
+	unsigned char dirty[4];
+
+	/* MicroP vsi */
+	union {
+		struct vdec_vp9_slice_init_vsi *init_vsi;
+		struct vdec_vp9_slice_vsi *vsi;
+	};
+	struct vdec_vp9_slice_vsi *core_vsi;
+
+	struct vdec_vp9_slice_pfc sc_pfc;
+	struct vdec_vp9_slice_counts_map counts_map;
+	struct v4l2_vp9_frame_symbol_counts counts_helper;
+};
+
+/*
+ * all VP9 instances could share this default frame context.
+ */
+static struct vdec_vp9_slice_frame_ctx *vdec_vp9_slice_default_frame_ctx;
+static DEFINE_MUTEX(vdec_vp9_slice_frame_ctx_lock);
+
+static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf);
+
+static int vdec_vp9_slice_init_default_frame_ctx(struct vdec_vp9_slice_instance *instance)
+{
+	struct vdec_vp9_slice_frame_ctx *remote_frame_ctx;
+	struct vdec_vp9_slice_frame_ctx *frame_ctx;
+	struct mtk_vcodec_dec_ctx *ctx;
+	struct vdec_vp9_slice_init_vsi *vsi;
+	int ret = 0;
+
+	ctx = instance->ctx;
+	vsi = instance->vpu.vsi;
+	if (!ctx || !vsi)
+		return -EINVAL;
+
+	remote_frame_ctx = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler,
+						     (u32)vsi->default_frame_ctx);
+	if (!remote_frame_ctx) {
+		mtk_vdec_err(ctx, "failed to map default frame ctx\n");
+		return -EINVAL;
+	}
+
+	mutex_lock(&vdec_vp9_slice_frame_ctx_lock);
+	if (vdec_vp9_slice_default_frame_ctx)
+		goto out;
+
+	frame_ctx = kmemdup(remote_frame_ctx, sizeof(*frame_ctx), GFP_KERNEL);
+	if (!frame_ctx) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	vdec_vp9_slice_default_frame_ctx = frame_ctx;
+
+out:
+	mutex_unlock(&vdec_vp9_slice_frame_ctx_lock);
+
+	return ret;
+}
+
+static int vdec_vp9_slice_alloc_working_buffer(struct vdec_vp9_slice_instance *instance,
+					       struct vdec_vp9_slice_vsi *vsi)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	enum vdec_vp9_slice_resolution_level level;
+	/* super blocks */
+	unsigned int max_sb_w;
+	unsigned int max_sb_h;
+	unsigned int max_w;
+	unsigned int max_h;
+	unsigned int w;
+	unsigned int h;
+	size_t size;
+	int ret;
+	int i;
+
+	w = vsi->frame.uh.frame_width;
+	h = vsi->frame.uh.frame_height;
+
+	if (w > VCODEC_DEC_4K_CODED_WIDTH ||
+	    h > VCODEC_DEC_4K_CODED_HEIGHT) {
+		return -EINVAL;
+	} else if (w > MTK_VDEC_MAX_W || h > MTK_VDEC_MAX_H) {
+		/* 4K */
+		level = VP9_RES_4K;
+		max_w = VCODEC_DEC_4K_CODED_WIDTH;
+		max_h = VCODEC_DEC_4K_CODED_HEIGHT;
+	} else {
+		/* FHD */
+		level = VP9_RES_FHD;
+		max_w = MTK_VDEC_MAX_W;
+		max_h = MTK_VDEC_MAX_H;
+	}
+
+	if (level == instance->level)
+		return 0;
+
+	mtk_vdec_debug(ctx, "resolution level changed, from %u to %u, %ux%u",
+		       instance->level, level, w, h);
+
+	max_sb_w = DIV_ROUND_UP(max_w, 64);
+	max_sb_h = DIV_ROUND_UP(max_h, 64);
+	ret = -ENOMEM;
+
+	/*
+	 * Lat-flush must wait core idle, otherwise core will
+	 * use released buffers
+	 */
+
+	size = (max_sb_w * max_sb_h + 2) * 576;
+	for (i = 0; i < 2; i++) {
+		if (instance->mv[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->mv[i]);
+		instance->mv[i].size = size;
+		if (mtk_vcodec_mem_alloc(ctx, &instance->mv[i]))
+			goto err;
+	}
+
+	size = (max_sb_w * max_sb_h * 32) + 256;
+	for (i = 0; i < 2; i++) {
+		if (instance->seg[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->seg[i]);
+		instance->seg[i].size = size;
+		if (mtk_vcodec_mem_alloc(ctx, &instance->seg[i]))
+			goto err;
+	}
+
+	if (!instance->tile.va) {
+		instance->tile.size = VP9_TILE_BUF_SIZE;
+		if (mtk_vcodec_mem_alloc(ctx, &instance->tile))
+			goto err;
+	}
+
+	if (!instance->prob.va) {
+		instance->prob.size = VP9_PROB_BUF_SIZE;
+		if (mtk_vcodec_mem_alloc(ctx, &instance->prob))
+			goto err;
+	}
+
+	if (!instance->counts.va) {
+		instance->counts.size = VP9_COUNTS_BUF_SIZE;
+		if (mtk_vcodec_mem_alloc(ctx, &instance->counts))
+			goto err;
+	}
+
+	instance->level = level;
+	return 0;
+
+err:
+	instance->level = VP9_RES_NONE;
+	return ret;
+}
+
+static void vdec_vp9_slice_free_working_buffer(struct vdec_vp9_slice_instance *instance)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(instance->mv); i++) {
+		if (instance->mv[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->mv[i]);
+	}
+	for (i = 0; i < ARRAY_SIZE(instance->seg); i++) {
+		if (instance->seg[i].va)
+			mtk_vcodec_mem_free(ctx, &instance->seg[i]);
+	}
+	if (instance->tile.va)
+		mtk_vcodec_mem_free(ctx, &instance->tile);
+	if (instance->prob.va)
+		mtk_vcodec_mem_free(ctx, &instance->prob);
+	if (instance->counts.va)
+		mtk_vcodec_mem_free(ctx, &instance->counts);
+
+	instance->level = VP9_RES_NONE;
+}
+
+static void vdec_vp9_slice_vsi_from_remote(struct vdec_vp9_slice_vsi *vsi,
+					   struct vdec_vp9_slice_vsi *remote_vsi,
+					   int skip)
+{
+	struct vdec_vp9_slice_frame *rf;
+	struct vdec_vp9_slice_frame *f;
+
+	/*
+	 * compressed header
+	 * dequant
+	 * buffer position
+	 * decode state
+	 */
+	if (!skip) {
+		rf = &remote_vsi->frame;
+		f = &vsi->frame;
+		memcpy(&f->ch, &rf->ch, sizeof(f->ch));
+		memcpy(&f->uh.dequant, &rf->uh.dequant, sizeof(f->uh.dequant));
+		memcpy(&vsi->trans, &remote_vsi->trans, sizeof(vsi->trans));
+	}
+
+	memcpy(&vsi->state, &remote_vsi->state, sizeof(vsi->state));
+}
+
+static void vdec_vp9_slice_vsi_to_remote(struct vdec_vp9_slice_vsi *vsi,
+					 struct vdec_vp9_slice_vsi *remote_vsi)
+{
+	memcpy(remote_vsi, vsi, sizeof(*vsi));
+}
+
+static int vdec_vp9_slice_tile_offset(int idx, int mi_num, int tile_log2)
+{
+	int sbs = (mi_num + 7) >> 3;
+	int offset = ((idx * sbs) >> tile_log2) << 3;
+
+	return min(offset, mi_num);
+}
+
+static
+int vdec_vp9_slice_setup_single_from_src_to_dst(struct vdec_vp9_slice_instance *instance)
+{
+	struct vb2_v4l2_buffer *src;
+	struct vb2_v4l2_buffer *dst;
+
+	src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx);
+	if (!src)
+		return -EINVAL;
+
+	dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx);
+	if (!dst)
+		return -EINVAL;
+
+	v4l2_m2m_buf_copy_metadata(src, dst, true);
+
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_lat_from_src_buf(struct vdec_vp9_slice_instance *instance,
+						 struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_v4l2_buffer *src;
+	struct vb2_v4l2_buffer *dst;
+
+	src = v4l2_m2m_next_src_buf(instance->ctx->m2m_ctx);
+	if (!src)
+		return -EINVAL;
+
+	lat_buf->src_buf_req = src->vb2_buf.req_obj.req;
+
+	dst = &lat_buf->ts_info;
+	v4l2_m2m_buf_copy_metadata(src, dst, true);
+	return 0;
+}
+
+static void vdec_vp9_slice_setup_hdr(struct vdec_vp9_slice_instance *instance,
+				     struct vdec_vp9_slice_uncompressed_header *uh,
+				     struct v4l2_ctrl_vp9_frame *hdr)
+{
+	int i;
+
+	uh->profile = hdr->profile;
+	uh->last_frame_type = instance->frame_type;
+	uh->frame_type = !HDR_FLAG(KEY_FRAME);
+	uh->last_show_frame = instance->show_frame;
+	uh->show_frame = HDR_FLAG(SHOW_FRAME);
+	uh->error_resilient_mode = HDR_FLAG(ERROR_RESILIENT);
+	uh->bit_depth = hdr->bit_depth;
+	uh->last_frame_width = instance->width;
+	uh->last_frame_height = instance->height;
+	uh->frame_width = hdr->frame_width_minus_1 + 1;
+	uh->frame_height = hdr->frame_height_minus_1 + 1;
+	uh->intra_only = HDR_FLAG(INTRA_ONLY);
+	/* map v4l2 enum to values defined in VP9 spec for firmware */
+	switch (hdr->reset_frame_context) {
+	case V4L2_VP9_RESET_FRAME_CTX_NONE:
+		uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_NONE0;
+		break;
+	case V4L2_VP9_RESET_FRAME_CTX_SPEC:
+		uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_SPEC;
+		break;
+	case V4L2_VP9_RESET_FRAME_CTX_ALL:
+		uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_ALL;
+		break;
+	default:
+		uh->reset_frame_context = VP9_RESET_FRAME_CONTEXT_NONE0;
+		break;
+	}
+	/*
+	 * ref_frame_sign_bias specifies the intended direction
+	 * of the motion vector in time for each reference frame.
+	 * - INTRA_FRAME = 0,
+	 * - LAST_FRAME = 1,
+	 * - GOLDEN_FRAME = 2,
+	 * - ALTREF_FRAME = 3,
+	 * ref_frame_sign_bias[INTRA_FRAME] is always 0
+	 * and VDA only passes another 3 directions
+	 */
+	uh->ref_frame_sign_bias[0] = 0;
+	for (i = 0; i < 3; i++)
+		uh->ref_frame_sign_bias[i + 1] =
+			!!(hdr->ref_frame_sign_bias & (1 << i));
+	uh->allow_high_precision_mv = HDR_FLAG(ALLOW_HIGH_PREC_MV);
+	uh->interpolation_filter = hdr->interpolation_filter;
+	uh->refresh_frame_context = HDR_FLAG(REFRESH_FRAME_CTX);
+	uh->frame_parallel_decoding_mode = HDR_FLAG(PARALLEL_DEC_MODE);
+	uh->frame_context_idx = hdr->frame_context_idx;
+
+	/* tile info */
+	uh->tile_cols_log2 = hdr->tile_cols_log2;
+	uh->tile_rows_log2 = hdr->tile_rows_log2;
+
+	uh->uncompressed_header_size = hdr->uncompressed_header_size;
+	uh->header_size_in_bytes = hdr->compressed_header_size;
+}
+
+static void vdec_vp9_slice_setup_frame_ctx(struct vdec_vp9_slice_instance *instance,
+					   struct vdec_vp9_slice_uncompressed_header *uh,
+					   struct v4l2_ctrl_vp9_frame *hdr)
+{
+	int error_resilient_mode;
+	int reset_frame_context;
+	int key_frame;
+	int intra_only;
+	int i;
+
+	key_frame = HDR_FLAG(KEY_FRAME);
+	intra_only = HDR_FLAG(INTRA_ONLY);
+	error_resilient_mode = HDR_FLAG(ERROR_RESILIENT);
+	reset_frame_context = uh->reset_frame_context;
+
+	/*
+	 * according to "6.2 Uncompressed header syntax" in
+	 * "VP9 Bitstream & Decoding Process Specification",
+	 * reset @frame_context_idx when (FrameIsIntra || error_resilient_mode)
+	 */
+	if (key_frame || intra_only || error_resilient_mode) {
+		/*
+		 * @reset_frame_context specifies
+		 * whether the frame context should be
+		 * reset to default values:
+		 * 0 or 1 means do not reset any frame context
+		 * 2 resets just the context specified in the frame header
+		 * 3 resets all contexts
+		 */
+		if (key_frame || error_resilient_mode ||
+		    reset_frame_context == 3) {
+			/* use default table */
+			for (i = 0; i < 4; i++)
+				instance->dirty[i] = 0;
+		} else if (reset_frame_context == 2) {
+			instance->dirty[uh->frame_context_idx] = 0;
+		}
+		uh->frame_context_idx = 0;
+	}
+}
+
+static void vdec_vp9_slice_setup_loop_filter(struct vdec_vp9_slice_uncompressed_header *uh,
+					     struct v4l2_vp9_loop_filter *lf)
+{
+	int i;
+
+	uh->loop_filter_level = lf->level;
+	uh->loop_filter_sharpness = lf->sharpness;
+	uh->loop_filter_delta_enabled = LF_FLAG(DELTA_ENABLED);
+	for (i = 0; i < 4; i++)
+		uh->loop_filter_ref_deltas[i] = lf->ref_deltas[i];
+	for (i = 0; i < 2; i++)
+		uh->loop_filter_mode_deltas[i] = lf->mode_deltas[i];
+}
+
+static void vdec_vp9_slice_setup_quantization(struct vdec_vp9_slice_uncompressed_header *uh,
+					      struct v4l2_vp9_quantization *quant)
+{
+	uh->base_q_idx = quant->base_q_idx;
+	uh->delta_q_y_dc = quant->delta_q_y_dc;
+	uh->delta_q_uv_dc = quant->delta_q_uv_dc;
+	uh->delta_q_uv_ac = quant->delta_q_uv_ac;
+}
+
+static void vdec_vp9_slice_setup_segmentation(struct vdec_vp9_slice_uncompressed_header *uh,
+					      struct v4l2_vp9_segmentation *seg)
+{
+	int i;
+	int j;
+
+	uh->segmentation_enabled = SEG_FLAG(ENABLED);
+	uh->segmentation_update_map = SEG_FLAG(UPDATE_MAP);
+	for (i = 0; i < 7; i++)
+		uh->segmentation_tree_probs[i] = seg->tree_probs[i];
+	uh->segmentation_temporal_udpate = SEG_FLAG(TEMPORAL_UPDATE);
+	for (i = 0; i < 3; i++)
+		uh->segmentation_pred_prob[i] = seg->pred_probs[i];
+	uh->segmentation_update_data = SEG_FLAG(UPDATE_DATA);
+	uh->segmentation_abs_or_delta_update = SEG_FLAG(ABS_OR_DELTA_UPDATE);
+	for (i = 0; i < 8; i++) {
+		uh->feature_enabled[i] = seg->feature_enabled[i];
+		for (j = 0; j < 4; j++)
+			uh->feature_value[i][j] = seg->feature_data[i][j];
+	}
+}
+
+static int vdec_vp9_slice_setup_tile(struct vdec_vp9_slice_vsi *vsi,
+				     struct v4l2_ctrl_vp9_frame *hdr)
+{
+	unsigned int rows_log2;
+	unsigned int cols_log2;
+	unsigned int rows;
+	unsigned int cols;
+	unsigned int mi_rows;
+	unsigned int mi_cols;
+	struct vdec_vp9_slice_tiles *tiles;
+	int offset;
+	int start;
+	int end;
+	int i;
+
+	rows_log2 = hdr->tile_rows_log2;
+	cols_log2 = hdr->tile_cols_log2;
+	rows = 1 << rows_log2;
+	cols = 1 << cols_log2;
+	tiles = &vsi->frame.tiles;
+	tiles->actual_rows = 0;
+
+	if (rows > 4 || cols > 64)
+		return -EINVAL;
+
+	/* setup mi rows/cols information */
+	mi_rows = (hdr->frame_height_minus_1 + 1 + 7) >> 3;
+	mi_cols = (hdr->frame_width_minus_1 + 1 + 7) >> 3;
+
+	for (i = 0; i < rows; i++) {
+		start = vdec_vp9_slice_tile_offset(i, mi_rows, rows_log2);
+		end = vdec_vp9_slice_tile_offset(i + 1, mi_rows, rows_log2);
+		offset = end - start;
+		tiles->mi_rows[i] = (offset + 7) >> 3;
+		if (tiles->mi_rows[i])
+			tiles->actual_rows++;
+	}
+
+	for (i = 0; i < cols; i++) {
+		start = vdec_vp9_slice_tile_offset(i, mi_cols, cols_log2);
+		end = vdec_vp9_slice_tile_offset(i + 1, mi_cols, cols_log2);
+		offset = end - start;
+		tiles->mi_cols[i] = (offset + 7) >> 3;
+	}
+
+	return 0;
+}
+
+static void vdec_vp9_slice_setup_state(struct vdec_vp9_slice_vsi *vsi)
+{
+	memset(&vsi->state, 0, sizeof(vsi->state));
+}
+
+static void vdec_vp9_slice_setup_ref_idx(struct vdec_vp9_slice_pfc *pfc,
+					 struct v4l2_ctrl_vp9_frame *hdr)
+{
+	pfc->ref_idx[0] = hdr->last_frame_ts;
+	pfc->ref_idx[1] = hdr->golden_frame_ts;
+	pfc->ref_idx[2] = hdr->alt_frame_ts;
+}
+
+static int vdec_vp9_slice_setup_pfc(struct vdec_vp9_slice_instance *instance,
+				    struct vdec_vp9_slice_pfc *pfc)
+{
+	struct v4l2_ctrl_vp9_frame *hdr;
+	struct vdec_vp9_slice_uncompressed_header *uh;
+	struct v4l2_ctrl *hdr_ctrl;
+	struct vdec_vp9_slice_vsi *vsi;
+	int ret;
+
+	/* frame header */
+	hdr_ctrl = v4l2_ctrl_find(&instance->ctx->ctrl_hdl, V4L2_CID_STATELESS_VP9_FRAME);
+	if (!hdr_ctrl || !hdr_ctrl->p_cur.p)
+		return -EINVAL;
+
+	hdr = hdr_ctrl->p_cur.p;
+	vsi = &pfc->vsi;
+	uh = &vsi->frame.uh;
+
+	/* setup vsi information */
+	vdec_vp9_slice_setup_hdr(instance, uh, hdr);
+	vdec_vp9_slice_setup_frame_ctx(instance, uh, hdr);
+	vdec_vp9_slice_setup_loop_filter(uh, &hdr->lf);
+	vdec_vp9_slice_setup_quantization(uh, &hdr->quant);
+	vdec_vp9_slice_setup_segmentation(uh, &hdr->seg);
+	ret = vdec_vp9_slice_setup_tile(vsi, hdr);
+	if (ret)
+		return ret;
+	vdec_vp9_slice_setup_state(vsi);
+
+	/* core stage needs buffer index to get ref y/c ... */
+	vdec_vp9_slice_setup_ref_idx(pfc, hdr);
+
+	pfc->seq = instance->seq;
+	instance->seq++;
+
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_lat_buffer(struct vdec_vp9_slice_instance *instance,
+					   struct vdec_vp9_slice_vsi *vsi,
+					   struct mtk_vcodec_mem *bs,
+					   struct vdec_lat_buf *lat_buf)
+{
+	int i;
+
+	vsi->bs.buf.dma_addr = bs->dma_addr;
+	vsi->bs.buf.size = bs->size;
+	vsi->bs.frame.dma_addr = bs->dma_addr;
+	vsi->bs.frame.size = bs->size;
+
+	for (i = 0; i < 2; i++) {
+		vsi->mv[i].dma_addr = instance->mv[i].dma_addr;
+		vsi->mv[i].size = instance->mv[i].size;
+	}
+	for (i = 0; i < 2; i++) {
+		vsi->seg[i].dma_addr = instance->seg[i].dma_addr;
+		vsi->seg[i].size = instance->seg[i].size;
+	}
+	vsi->tile.dma_addr = instance->tile.dma_addr;
+	vsi->tile.size = instance->tile.size;
+	vsi->prob.dma_addr = instance->prob.dma_addr;
+	vsi->prob.size = instance->prob.size;
+	vsi->counts.dma_addr = instance->counts.dma_addr;
+	vsi->counts.size = instance->counts.size;
+
+	vsi->ube.dma_addr = lat_buf->ctx->msg_queue.wdma_addr.dma_addr;
+	vsi->ube.size = lat_buf->ctx->msg_queue.wdma_addr.size;
+	vsi->trans.dma_addr = lat_buf->ctx->msg_queue.wdma_wptr_addr;
+	/* used to store trans end */
+	vsi->trans.dma_addr_end = lat_buf->ctx->msg_queue.wdma_rptr_addr;
+	vsi->err_map.dma_addr = lat_buf->wdma_err_addr.dma_addr;
+	vsi->err_map.size = lat_buf->wdma_err_addr.size;
+
+	vsi->row_info.buf = 0;
+	vsi->row_info.size = 0;
+
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_prob_buffer(struct vdec_vp9_slice_instance *instance,
+					    struct vdec_vp9_slice_vsi *vsi)
+{
+	struct vdec_vp9_slice_frame_ctx *frame_ctx;
+	struct vdec_vp9_slice_uncompressed_header *uh;
+
+	uh = &vsi->frame.uh;
+
+	mtk_vdec_debug(instance->ctx, "ctx dirty %u idx %d\n",
+		       instance->dirty[uh->frame_context_idx],
+		       uh->frame_context_idx);
+
+	if (instance->dirty[uh->frame_context_idx])
+		frame_ctx = &instance->frame_ctx[uh->frame_context_idx];
+	else
+		frame_ctx = vdec_vp9_slice_default_frame_ctx;
+	memcpy(instance->prob.va, frame_ctx, sizeof(*frame_ctx));
+
+	return 0;
+}
+
+static void vdec_vp9_slice_setup_seg_buffer(struct vdec_vp9_slice_instance *instance,
+					    struct vdec_vp9_slice_vsi *vsi,
+					    struct mtk_vcodec_mem *buf)
+{
+	struct vdec_vp9_slice_uncompressed_header *uh;
+
+	/* reset segment buffer */
+	uh = &vsi->frame.uh;
+	if (uh->frame_type == 0 ||
+	    uh->intra_only ||
+	    uh->error_resilient_mode ||
+	    uh->frame_width != instance->width ||
+	    uh->frame_height != instance->height) {
+		mtk_vdec_debug(instance->ctx, "reset seg\n");
+		memset(buf->va, 0, buf->size);
+	}
+}
+
+/*
+ * parse tiles according to `6.4 Decode tiles syntax`
+ * in "vp9-bitstream-specification"
+ *
+ * frame contains uncompress header, compressed header and several tiles.
+ * this function parses tiles' position and size, stores them to tile buffer
+ * for decoding.
+ */
+static int vdec_vp9_slice_setup_tile_buffer(struct vdec_vp9_slice_instance *instance,
+					    struct vdec_vp9_slice_vsi *vsi,
+					    struct mtk_vcodec_mem *bs)
+{
+	struct vdec_vp9_slice_uncompressed_header *uh;
+	unsigned int rows_log2;
+	unsigned int cols_log2;
+	unsigned int rows;
+	unsigned int cols;
+	unsigned int mi_row;
+	unsigned int mi_col;
+	unsigned int offset;
+	unsigned int pa;
+	unsigned int size;
+	struct vdec_vp9_slice_tiles *tiles;
+	unsigned char *pos;
+	unsigned char *end;
+	unsigned char *va;
+	unsigned int *tb;
+	int i;
+	int j;
+
+	uh = &vsi->frame.uh;
+	rows_log2 = uh->tile_rows_log2;
+	cols_log2 = uh->tile_cols_log2;
+	rows = 1 << rows_log2;
+	cols = 1 << cols_log2;
+
+	if (rows > 4 || cols > 64) {
+		mtk_vdec_err(instance->ctx, "tile_rows %u tile_cols %u\n", rows, cols);
+		return -EINVAL;
+	}
+
+	offset = uh->uncompressed_header_size +
+		uh->header_size_in_bytes;
+	if (bs->size <= offset) {
+		mtk_vdec_err(instance->ctx, "bs size %zu tile offset %u\n", bs->size, offset);
+		return -EINVAL;
+	}
+
+	tiles = &vsi->frame.tiles;
+	/* setup tile buffer */
+
+	va = (unsigned char *)bs->va;
+	pos = va + offset;
+	end = va + bs->size;
+	/* truncated */
+	pa = (unsigned int)bs->dma_addr + offset;
+	tb = instance->tile.va;
+	for (i = 0; i < rows; i++) {
+		for (j = 0; j < cols; j++) {
+			if (i == rows - 1 &&
+			    j == cols - 1) {
+				size = (unsigned int)(end - pos);
+			} else {
+				if (end - pos < 4)
+					return -EINVAL;
+
+				size = (pos[0] << 24) | (pos[1] << 16) |
+					(pos[2] << 8) | pos[3];
+				pos += 4;
+				pa += 4;
+				offset += 4;
+				if (end - pos < size)
+					return -EINVAL;
+			}
+			tiles->size[i][j] = size;
+			if (tiles->mi_rows[i]) {
+				*tb++ = (size << 3) + ((offset << 3) & 0x7f);
+				*tb++ = pa & ~0xf;
+				*tb++ = (pa << 3) & 0x7f;
+				mi_row = (tiles->mi_rows[i] - 1) & 0x1ff;
+				mi_col = (tiles->mi_cols[j] - 1) & 0x3f;
+				*tb++ = (mi_row << 6) + mi_col;
+			}
+			pos += size;
+			pa += size;
+			offset += size;
+		}
+	}
+
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_lat(struct vdec_vp9_slice_instance *instance,
+				    struct mtk_vcodec_mem *bs,
+				    struct vdec_lat_buf *lat_buf,
+				    struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi = &pfc->vsi;
+	int ret;
+
+	ret = vdec_vp9_slice_setup_lat_from_src_buf(instance, lat_buf);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_pfc(instance, pfc);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_alloc_working_buffer(instance, vsi);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_lat_buffer(instance, vsi, bs, lat_buf);
+	if (ret)
+		goto err;
+
+	vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[0]);
+
+	/* setup prob/tile buffers for LAT */
+
+	ret = vdec_vp9_slice_setup_prob_buffer(instance, vsi);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_tile_buffer(instance, vsi, bs);
+	if (ret)
+		goto err;
+
+	return 0;
+
+err:
+	return ret;
+}
+
+static
+void vdec_vp9_slice_map_counts_eob_coef(unsigned int i, unsigned int j, unsigned int k,
+					struct vdec_vp9_slice_frame_counts *counts,
+					struct v4l2_vp9_frame_symbol_counts *counts_helper)
+{
+	u32 l = 0, m;
+
+	/*
+	 * helper eo -> mtk eo
+	 * helpre e1 -> mtk c3
+	 * helper c0 -> c0
+	 * helper c1 -> c1
+	 * helper c2 -> c2
+	 */
+	for (m = 0; m < 3; m++) {
+		counts_helper->coeff[i][j][k][l][m] =
+			(u32 (*)[3]) & counts->coef_probs[i][j][k].band_0[m];
+		counts_helper->eob[i][j][k][l][m][0] =
+			&counts->eob_branch[i][j][k].band_0[m];
+		counts_helper->eob[i][j][k][l][m][1] =
+			&counts->coef_probs[i][j][k].band_0[m][3];
+	}
+
+	for (l = 1; l < 6; l++) {
+		for (m = 0; m < 6; m++) {
+			counts_helper->coeff[i][j][k][l][m] =
+				(u32 (*)[3]) & counts->coef_probs[i][j][k].band_1_5[l - 1][m];
+			counts_helper->eob[i][j][k][l][m][0] =
+				&counts->eob_branch[i][j][k].band_1_5[l - 1][m];
+			counts_helper->eob[i][j][k][l][m][1] =
+				&counts->coef_probs[i][j][k].band_1_5[l - 1][m][3];
+		}
+	}
+}
+
+static void vdec_vp9_slice_counts_map_helper(struct vdec_vp9_slice_counts_map *counts_map,
+					     struct vdec_vp9_slice_frame_counts *counts,
+					     struct v4l2_vp9_frame_symbol_counts *counts_helper)
+{
+	int i, j, k;
+
+	counts_helper->partition = &counts->partition;
+	counts_helper->intra_inter = &counts->intra_inter;
+	counts_helper->tx32p = &counts->tx_p32x32;
+	counts_helper->tx16p = &counts->tx_p16x16;
+	counts_helper->tx8p = &counts->tx_p8x8;
+	counts_helper->uv_mode = &counts->uv_mode;
+
+	counts_helper->comp = &counts->comp_inter;
+	counts_helper->comp_ref = &counts->comp_ref;
+	counts_helper->single_ref = &counts->single_ref;
+	counts_helper->mv_mode = &counts->inter_mode;
+	counts_helper->mv_joint = &counts->joint;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->skip); i++)
+		memcpy(counts_map->skip[i], counts->skip[i],
+		       sizeof(counts_map->skip[0]));
+	counts_helper->skip = &counts_map->skip;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->y_mode); i++)
+		memcpy(counts_map->y_mode[i], counts->y_mode[i],
+		       sizeof(counts_map->y_mode[0]));
+	counts_helper->y_mode = &counts_map->y_mode;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->filter); i++)
+		memcpy(counts_map->filter[i], counts->switchable_interp[i],
+		       sizeof(counts_map->filter[0]));
+	counts_helper->filter = &counts_map->filter;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->sign); i++)
+		memcpy(counts_map->sign[i], counts->mvcomp[i].sign,
+		       sizeof(counts_map->sign[0]));
+	counts_helper->sign = &counts_map->sign;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->classes); i++)
+		memcpy(counts_map->classes[i], counts->mvcomp[i].classes,
+		       sizeof(counts_map->classes[0]));
+	counts_helper->classes = &counts_map->classes;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->class0); i++)
+		memcpy(counts_map->class0[i], counts->mvcomp[i].class0,
+		       sizeof(counts_map->class0[0]));
+	counts_helper->class0 = &counts_map->class0;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->bits); i++)
+		for (j = 0; j < ARRAY_SIZE(counts_map->bits[0]); j++)
+			memcpy(counts_map->bits[i][j], counts->mvcomp[i].bits[j],
+			       sizeof(counts_map->bits[0][0]));
+	counts_helper->bits = &counts_map->bits;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->class0_fp); i++)
+		for (j = 0; j < ARRAY_SIZE(counts_map->class0_fp[0]); j++)
+			memcpy(counts_map->class0_fp[i][j], counts->mvcomp[i].class0_fp[j],
+			       sizeof(counts_map->class0_fp[0][0]));
+	counts_helper->class0_fp = &counts_map->class0_fp;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->fp); i++)
+		memcpy(counts_map->fp[i], counts->mvcomp[i].fp,
+		       sizeof(counts_map->fp[0]));
+	counts_helper->fp = &counts_map->fp;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->class0_hp); i++)
+		memcpy(counts_map->class0_hp[i], counts->mvcomp[i].class0_hp,
+		       sizeof(counts_map->class0_hp[0]));
+	counts_helper->class0_hp = &counts_map->class0_hp;
+
+	for (i = 0; i < ARRAY_SIZE(counts_map->hp); i++)
+		memcpy(counts_map->hp[i], counts->mvcomp[i].hp, sizeof(counts_map->hp[0]));
+
+	counts_helper->hp = &counts_map->hp;
+
+	for (i = 0; i < 4; i++)
+		for (j = 0; j < 2; j++)
+			for (k = 0; k < 2; k++)
+				vdec_vp9_slice_map_counts_eob_coef(i, j, k, counts, counts_helper);
+}
+
+static void vdec_vp9_slice_map_to_coef(unsigned int i, unsigned int j, unsigned int k,
+				       struct vdec_vp9_slice_frame_ctx *frame_ctx,
+				       struct v4l2_vp9_frame_context *frame_ctx_helper)
+{
+	u32 l, m;
+
+	for (l = 0; l < ARRAY_SIZE(frame_ctx_helper->coef[0][0][0]); l++) {
+		for (m = 0; m < VP9_BAND_6(l); m++) {
+			memcpy(frame_ctx_helper->coef[i][j][k][l][m],
+			       frame_ctx->coef_probs[i][j][k][l].probs[m],
+			       sizeof(frame_ctx_helper->coef[i][j][k][l][0]));
+		}
+	}
+}
+
+static void vdec_vp9_slice_map_from_coef(unsigned int i, unsigned int j, unsigned int k,
+					 struct vdec_vp9_slice_frame_ctx *frame_ctx,
+					 struct v4l2_vp9_frame_context *frame_ctx_helper)
+{
+	u32 l, m;
+
+	for (l = 0; l < ARRAY_SIZE(frame_ctx_helper->coef[0][0][0]); l++) {
+		for (m = 0; m < VP9_BAND_6(l); m++) {
+			memcpy(frame_ctx->coef_probs[i][j][k][l].probs[m],
+			       frame_ctx_helper->coef[i][j][k][l][m],
+			       sizeof(frame_ctx_helper->coef[i][j][k][l][0]));
+		}
+	}
+}
+
+static
+void vdec_vp9_slice_framectx_map_helper(bool frame_is_intra,
+					struct vdec_vp9_slice_frame_ctx *pre_frame_ctx,
+					struct vdec_vp9_slice_frame_ctx *frame_ctx,
+					struct v4l2_vp9_frame_context *frame_ctx_helper)
+{
+	struct v4l2_vp9_frame_mv_context *mv = &frame_ctx_helper->mv;
+	u32 i, j, k;
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->coef); i++)
+		for (j = 0; j < ARRAY_SIZE(frame_ctx_helper->coef[0]); j++)
+			for (k = 0; k < ARRAY_SIZE(frame_ctx_helper->coef[0][0]); k++)
+				vdec_vp9_slice_map_to_coef(i, j, k, pre_frame_ctx,
+							   frame_ctx_helper);
+
+	/*
+	 * use previous prob when frame is not intra or
+	 * we should use the prob updated by the compressed header parse
+	 */
+	if (!frame_is_intra)
+		frame_ctx = pre_frame_ctx;
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx8); i++)
+		memcpy(frame_ctx_helper->tx8[i], frame_ctx->tx_p8x8[i],
+		       sizeof(frame_ctx_helper->tx8[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx16); i++)
+		memcpy(frame_ctx_helper->tx16[i], frame_ctx->tx_p16x16[i],
+		       sizeof(frame_ctx_helper->tx16[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx32); i++)
+		memcpy(frame_ctx_helper->tx32[i], frame_ctx->tx_p32x32[i],
+		       sizeof(frame_ctx_helper->tx32[0]));
+
+	memcpy(frame_ctx_helper->skip, frame_ctx->skip_probs, sizeof(frame_ctx_helper->skip));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->inter_mode); i++)
+		memcpy(frame_ctx_helper->inter_mode[i], frame_ctx->inter_mode_probs[i],
+		       sizeof(frame_ctx_helper->inter_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->interp_filter); i++)
+		memcpy(frame_ctx_helper->interp_filter[i], frame_ctx->switch_interp_prob[i],
+		       sizeof(frame_ctx_helper->interp_filter[0]));
+
+	memcpy(frame_ctx_helper->is_inter, frame_ctx->intra_inter_prob,
+	       sizeof(frame_ctx_helper->is_inter));
+
+	memcpy(frame_ctx_helper->comp_mode, frame_ctx->comp_inter_prob,
+	       sizeof(frame_ctx_helper->comp_mode));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->single_ref); i++)
+		memcpy(frame_ctx_helper->single_ref[i], frame_ctx->single_ref_prob[i],
+		       sizeof(frame_ctx_helper->single_ref[0]));
+
+	memcpy(frame_ctx_helper->comp_ref, frame_ctx->comp_ref_prob,
+	       sizeof(frame_ctx_helper->comp_ref));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->y_mode); i++)
+		memcpy(frame_ctx_helper->y_mode[i], frame_ctx->y_mode_prob[i],
+		       sizeof(frame_ctx_helper->y_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->uv_mode); i++)
+		memcpy(frame_ctx_helper->uv_mode[i], frame_ctx->uv_mode_prob[i],
+		       sizeof(frame_ctx_helper->uv_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->partition); i++)
+		memcpy(frame_ctx_helper->partition[i], frame_ctx->partition_prob[i],
+		       sizeof(frame_ctx_helper->partition[0]));
+
+	memcpy(mv->joint, frame_ctx->joint, sizeof(mv->joint));
+
+	for (i = 0; i < ARRAY_SIZE(mv->sign); i++)
+		mv->sign[i] = frame_ctx->sign_classes[i].sign;
+
+	for (i = 0; i < ARRAY_SIZE(mv->classes); i++)
+		memcpy(mv->classes[i], frame_ctx->sign_classes[i].classes,
+		       sizeof(mv->classes[i]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_bit); i++)
+		mv->class0_bit[i] = frame_ctx->class0_bits[i].class0[0];
+
+	for (i = 0; i < ARRAY_SIZE(mv->bits); i++)
+		memcpy(mv->bits[i], frame_ctx->class0_bits[i].bits, sizeof(mv->bits[0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_fr); i++)
+		for (j = 0; j < ARRAY_SIZE(mv->class0_fr[0]); j++)
+			memcpy(mv->class0_fr[i][j], frame_ctx->class0_fp_hp[i].class0_fp[j],
+			       sizeof(mv->class0_fr[0][0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->fr); i++)
+		memcpy(mv->fr[i], frame_ctx->class0_fp_hp[i].fp, sizeof(mv->fr[0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_hp); i++)
+		mv->class0_hp[i] = frame_ctx->class0_fp_hp[i].class0_hp;
+
+	for (i = 0; i < ARRAY_SIZE(mv->hp); i++)
+		mv->hp[i] = frame_ctx->class0_fp_hp[i].hp;
+}
+
+static void vdec_vp9_slice_helper_map_framectx(struct v4l2_vp9_frame_context *frame_ctx_helper,
+					       struct vdec_vp9_slice_frame_ctx *frame_ctx)
+{
+	struct v4l2_vp9_frame_mv_context *mv = &frame_ctx_helper->mv;
+	u32 i, j, k;
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx8); i++)
+		memcpy(frame_ctx->tx_p8x8[i], frame_ctx_helper->tx8[i],
+		       sizeof(frame_ctx_helper->tx8[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx16); i++)
+		memcpy(frame_ctx->tx_p16x16[i], frame_ctx_helper->tx16[i],
+		       sizeof(frame_ctx_helper->tx16[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->tx32); i++)
+		memcpy(frame_ctx->tx_p32x32[i], frame_ctx_helper->tx32[i],
+		       sizeof(frame_ctx_helper->tx32[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->coef); i++)
+		for (j = 0; j < ARRAY_SIZE(frame_ctx_helper->coef[0]); j++)
+			for (k = 0; k < ARRAY_SIZE(frame_ctx_helper->coef[0][0]); k++)
+				vdec_vp9_slice_map_from_coef(i, j, k, frame_ctx,
+							     frame_ctx_helper);
+
+	memcpy(frame_ctx->skip_probs, frame_ctx_helper->skip, sizeof(frame_ctx_helper->skip));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->inter_mode); i++)
+		memcpy(frame_ctx->inter_mode_probs[i], frame_ctx_helper->inter_mode[i],
+		       sizeof(frame_ctx_helper->inter_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->interp_filter); i++)
+		memcpy(frame_ctx->switch_interp_prob[i], frame_ctx_helper->interp_filter[i],
+		       sizeof(frame_ctx_helper->interp_filter[0]));
+
+	memcpy(frame_ctx->intra_inter_prob, frame_ctx_helper->is_inter,
+	       sizeof(frame_ctx_helper->is_inter));
+
+	memcpy(frame_ctx->comp_inter_prob, frame_ctx_helper->comp_mode,
+	       sizeof(frame_ctx_helper->comp_mode));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->single_ref); i++)
+		memcpy(frame_ctx->single_ref_prob[i], frame_ctx_helper->single_ref[i],
+		       sizeof(frame_ctx_helper->single_ref[0]));
+
+	memcpy(frame_ctx->comp_ref_prob, frame_ctx_helper->comp_ref,
+	       sizeof(frame_ctx_helper->comp_ref));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->y_mode); i++)
+		memcpy(frame_ctx->y_mode_prob[i], frame_ctx_helper->y_mode[i],
+		       sizeof(frame_ctx_helper->y_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->uv_mode); i++)
+		memcpy(frame_ctx->uv_mode_prob[i], frame_ctx_helper->uv_mode[i],
+		       sizeof(frame_ctx_helper->uv_mode[0]));
+
+	for (i = 0; i < ARRAY_SIZE(frame_ctx_helper->partition); i++)
+		memcpy(frame_ctx->partition_prob[i], frame_ctx_helper->partition[i],
+		       sizeof(frame_ctx_helper->partition[0]));
+
+	memcpy(frame_ctx->joint, mv->joint, sizeof(mv->joint));
+
+	for (i = 0; i < ARRAY_SIZE(mv->sign); i++)
+		frame_ctx->sign_classes[i].sign = mv->sign[i];
+
+	for (i = 0; i < ARRAY_SIZE(mv->classes); i++)
+		memcpy(frame_ctx->sign_classes[i].classes, mv->classes[i],
+		       sizeof(mv->classes[i]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_bit); i++)
+		frame_ctx->class0_bits[i].class0[0] = mv->class0_bit[i];
+
+	for (i = 0; i < ARRAY_SIZE(mv->bits); i++)
+		memcpy(frame_ctx->class0_bits[i].bits, mv->bits[i], sizeof(mv->bits[0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_fr); i++)
+		for (j = 0; j < ARRAY_SIZE(mv->class0_fr[0]); j++)
+			memcpy(frame_ctx->class0_fp_hp[i].class0_fp[j], mv->class0_fr[i][j],
+			       sizeof(mv->class0_fr[0][0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->fr); i++)
+		memcpy(frame_ctx->class0_fp_hp[i].fp, mv->fr[i], sizeof(mv->fr[0]));
+
+	for (i = 0; i < ARRAY_SIZE(mv->class0_hp); i++)
+		frame_ctx->class0_fp_hp[i].class0_hp = mv->class0_hp[i];
+
+	for (i = 0; i < ARRAY_SIZE(mv->hp); i++)
+		frame_ctx->class0_fp_hp[i].hp = mv->hp[i];
+}
+
+static int vdec_vp9_slice_update_prob(struct vdec_vp9_slice_instance *instance,
+				      struct vdec_vp9_slice_vsi *vsi)
+{
+	struct vdec_vp9_slice_frame_ctx *pre_frame_ctx;
+	struct v4l2_vp9_frame_context *pre_frame_ctx_helper;
+	struct vdec_vp9_slice_frame_ctx *frame_ctx;
+	struct vdec_vp9_slice_frame_counts *counts;
+	struct v4l2_vp9_frame_symbol_counts *counts_helper;
+	struct vdec_vp9_slice_uncompressed_header *uh;
+	bool frame_is_intra;
+	bool use_128;
+
+	uh = &vsi->frame.uh;
+	pre_frame_ctx = &instance->frame_ctx[uh->frame_context_idx];
+	pre_frame_ctx_helper = &instance->frame_ctx_helper;
+	frame_ctx = (struct vdec_vp9_slice_frame_ctx *)instance->prob.va;
+	counts = (struct vdec_vp9_slice_frame_counts *)instance->counts.va;
+	counts_helper = &instance->counts_helper;
+
+	if (!uh->refresh_frame_context)
+		return 0;
+
+	if (!uh->frame_parallel_decoding_mode) {
+		vdec_vp9_slice_counts_map_helper(&instance->counts_map, counts, counts_helper);
+
+		frame_is_intra = !vsi->frame.uh.frame_type || vsi->frame.uh.intra_only;
+		/* check default prob */
+		if (!instance->dirty[uh->frame_context_idx])
+			vdec_vp9_slice_framectx_map_helper(frame_is_intra,
+							   vdec_vp9_slice_default_frame_ctx,
+							   frame_ctx,
+							   pre_frame_ctx_helper);
+		else
+			vdec_vp9_slice_framectx_map_helper(frame_is_intra,
+							   pre_frame_ctx,
+							   frame_ctx,
+							   pre_frame_ctx_helper);
+
+		use_128 = !frame_is_intra && !vsi->frame.uh.last_frame_type;
+		v4l2_vp9_adapt_coef_probs(pre_frame_ctx_helper,
+					  counts_helper,
+					  use_128,
+					  frame_is_intra);
+		if (!frame_is_intra)
+			v4l2_vp9_adapt_noncoef_probs(pre_frame_ctx_helper,
+						     counts_helper,
+						     V4L2_VP9_REFERENCE_MODE_SINGLE_REFERENCE,
+						     vsi->frame.uh.interpolation_filter,
+						     vsi->frame.ch.tx_mode,
+						     vsi->frame.uh.allow_high_precision_mv ?
+						     V4L2_VP9_FRAME_FLAG_ALLOW_HIGH_PREC_MV : 0);
+		vdec_vp9_slice_helper_map_framectx(pre_frame_ctx_helper, pre_frame_ctx);
+	} else {
+		memcpy(pre_frame_ctx, frame_ctx, sizeof(*frame_ctx));
+	}
+
+	instance->dirty[uh->frame_context_idx] = 1;
+
+	return 0;
+}
+
+static int vdec_vp9_slice_update_single(struct vdec_vp9_slice_instance *instance,
+					struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi;
+
+	vsi = &pfc->vsi;
+	memcpy(&pfc->state[0], &vsi->state, sizeof(vsi->state));
+
+	mtk_vdec_debug(instance->ctx, "Frame %u Y_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[0], vsi->state.crc[1],
+		       vsi->state.crc[2], vsi->state.crc[3]);
+	mtk_vdec_debug(instance->ctx, "Frame %u C_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[4], vsi->state.crc[5],
+		       vsi->state.crc[6], vsi->state.crc[7]);
+
+	vdec_vp9_slice_update_prob(instance, vsi);
+
+	instance->width = vsi->frame.uh.frame_width;
+	instance->height = vsi->frame.uh.frame_height;
+	instance->frame_type = vsi->frame.uh.frame_type;
+	instance->show_frame = vsi->frame.uh.show_frame;
+
+	return 0;
+}
+
+static int vdec_vp9_slice_update_lat(struct vdec_vp9_slice_instance *instance,
+				     struct vdec_lat_buf *lat_buf,
+				     struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi;
+
+	vsi = &pfc->vsi;
+	memcpy(&pfc->state[0], &vsi->state, sizeof(vsi->state));
+
+	mtk_vdec_debug(instance->ctx, "Frame %u LAT CRC 0x%08x %lx %lx\n",
+		       pfc->seq, vsi->state.crc[0],
+		       (unsigned long)vsi->trans.dma_addr,
+		       (unsigned long)vsi->trans.dma_addr_end);
+
+	/* buffer full, need to re-decode */
+	if (vsi->state.full) {
+		/* buffer not enough */
+		if (vsi->trans.dma_addr_end - vsi->trans.dma_addr ==
+			vsi->ube.size)
+			return -ENOMEM;
+		return -EAGAIN;
+	}
+
+	vdec_vp9_slice_update_prob(instance, vsi);
+
+	instance->width = vsi->frame.uh.frame_width;
+	instance->height = vsi->frame.uh.frame_height;
+	instance->frame_type = vsi->frame.uh.frame_type;
+	instance->show_frame = vsi->frame.uh.show_frame;
+
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_core_to_dst_buf(struct vdec_vp9_slice_instance *instance,
+						struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_v4l2_buffer *dst;
+
+	dst = v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx);
+	if (!dst)
+		return -EINVAL;
+
+	v4l2_m2m_buf_copy_metadata(&lat_buf->ts_info, dst, true);
+	return 0;
+}
+
+static int vdec_vp9_slice_setup_core_buffer(struct vdec_vp9_slice_instance *instance,
+					    struct vdec_vp9_slice_pfc *pfc,
+					    struct vdec_vp9_slice_vsi *vsi,
+					    struct vdec_fb *fb,
+					    struct vdec_lat_buf *lat_buf)
+{
+	struct vb2_buffer *vb;
+	struct vb2_queue *vq;
+	struct vdec_vp9_slice_reference *ref;
+	int plane;
+	int size;
+	int w;
+	int h;
+	int i;
+
+	plane = instance->ctx->q_data[MTK_Q_DATA_DST].fmt->num_planes;
+	w = vsi->frame.uh.frame_width;
+	h = vsi->frame.uh.frame_height;
+	size = ALIGN(w, 64) * ALIGN(h, 64);
+
+	/* frame buffer */
+	vsi->fb.y.dma_addr = fb->base_y.dma_addr;
+	if (plane == 1)
+		vsi->fb.c.dma_addr = fb->base_y.dma_addr + size;
+	else
+		vsi->fb.c.dma_addr = fb->base_c.dma_addr;
+
+	/* reference buffers */
+	vq = v4l2_m2m_get_vq(instance->ctx->m2m_ctx,
+			     V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE);
+	if (!vq)
+		return -EINVAL;
+
+	/* get current output buffer */
+	vb = &v4l2_m2m_next_dst_buf(instance->ctx->m2m_ctx)->vb2_buf;
+	if (!vb)
+		return -EINVAL;
+
+	/* update internal buffer's width/height */
+	for (i = 0; i < vq->num_buffers; i++) {
+		if (vb == vq->bufs[i]) {
+			instance->dpb[i].width = w;
+			instance->dpb[i].height = h;
+			break;
+		}
+	}
+
+	/*
+	 * get buffer's width/height from instance
+	 * get buffer address from vb2buf
+	 */
+	for (i = 0; i < 3; i++) {
+		ref = &vsi->frame.ref[i];
+		vb = vb2_find_buffer(vq, pfc->ref_idx[i]);
+		if (!vb) {
+			ref->frame_width = w;
+			ref->frame_height = h;
+			memset(&vsi->ref[i], 0, sizeof(vsi->ref[i]));
+		} else {
+			int idx = vb->index;
+
+			ref->frame_width = instance->dpb[idx].width;
+			ref->frame_height = instance->dpb[idx].height;
+			vsi->ref[i].y.dma_addr =
+				vb2_dma_contig_plane_dma_addr(vb, 0);
+			if (plane == 1)
+				vsi->ref[i].c.dma_addr =
+					vsi->ref[i].y.dma_addr + size;
+			else
+				vsi->ref[i].c.dma_addr =
+					vb2_dma_contig_plane_dma_addr(vb, 1);
+		}
+	}
+
+	return 0;
+}
+
+static void vdec_vp9_slice_setup_single_buffer(struct vdec_vp9_slice_instance *instance,
+					       struct vdec_vp9_slice_pfc *pfc,
+					       struct vdec_vp9_slice_vsi *vsi,
+					       struct mtk_vcodec_mem *bs,
+					       struct vdec_fb *fb)
+{
+	int i;
+
+	vsi->bs.buf.dma_addr = bs->dma_addr;
+	vsi->bs.buf.size = bs->size;
+	vsi->bs.frame.dma_addr = bs->dma_addr;
+	vsi->bs.frame.size = bs->size;
+
+	for (i = 0; i < 2; i++) {
+		vsi->mv[i].dma_addr = instance->mv[i].dma_addr;
+		vsi->mv[i].size = instance->mv[i].size;
+	}
+	for (i = 0; i < 2; i++) {
+		vsi->seg[i].dma_addr = instance->seg[i].dma_addr;
+		vsi->seg[i].size = instance->seg[i].size;
+	}
+	vsi->tile.dma_addr = instance->tile.dma_addr;
+	vsi->tile.size = instance->tile.size;
+	vsi->prob.dma_addr = instance->prob.dma_addr;
+	vsi->prob.size = instance->prob.size;
+	vsi->counts.dma_addr = instance->counts.dma_addr;
+	vsi->counts.size = instance->counts.size;
+
+	vsi->row_info.buf = 0;
+	vsi->row_info.size = 0;
+
+	vdec_vp9_slice_setup_core_buffer(instance, pfc, vsi, fb, NULL);
+}
+
+static int vdec_vp9_slice_setup_core(struct vdec_vp9_slice_instance *instance,
+				     struct vdec_fb *fb,
+				     struct vdec_lat_buf *lat_buf,
+				     struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi = &pfc->vsi;
+	int ret;
+
+	vdec_vp9_slice_setup_state(vsi);
+
+	ret = vdec_vp9_slice_setup_core_to_dst_buf(instance, lat_buf);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_core_buffer(instance, pfc, vsi, fb, lat_buf);
+	if (ret)
+		goto err;
+
+	vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[1]);
+
+	return 0;
+
+err:
+	return ret;
+}
+
+static int vdec_vp9_slice_setup_single(struct vdec_vp9_slice_instance *instance,
+				       struct mtk_vcodec_mem *bs,
+				       struct vdec_fb *fb,
+				       struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi = &pfc->vsi;
+	int ret;
+
+	ret = vdec_vp9_slice_setup_single_from_src_to_dst(instance);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_pfc(instance, pfc);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_alloc_working_buffer(instance, vsi);
+	if (ret)
+		goto err;
+
+	vdec_vp9_slice_setup_single_buffer(instance, pfc, vsi, bs, fb);
+	vdec_vp9_slice_setup_seg_buffer(instance, vsi, &instance->seg[0]);
+
+	ret = vdec_vp9_slice_setup_prob_buffer(instance, vsi);
+	if (ret)
+		goto err;
+
+	ret = vdec_vp9_slice_setup_tile_buffer(instance, vsi, bs);
+	if (ret)
+		goto err;
+
+	return 0;
+
+err:
+	return ret;
+}
+
+static int vdec_vp9_slice_update_core(struct vdec_vp9_slice_instance *instance,
+				      struct vdec_lat_buf *lat_buf,
+				      struct vdec_vp9_slice_pfc *pfc)
+{
+	struct vdec_vp9_slice_vsi *vsi;
+
+	vsi = &pfc->vsi;
+	memcpy(&pfc->state[1], &vsi->state, sizeof(vsi->state));
+
+	mtk_vdec_debug(instance->ctx, "Frame %u Y_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[0], vsi->state.crc[1],
+		       vsi->state.crc[2], vsi->state.crc[3]);
+	mtk_vdec_debug(instance->ctx, "Frame %u C_CRC %08x %08x %08x %08x\n",
+		       pfc->seq, vsi->state.crc[4], vsi->state.crc[5],
+		       vsi->state.crc[6], vsi->state.crc[7]);
+
+	return 0;
+}
+
+static int vdec_vp9_slice_init(struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_vp9_slice_instance *instance;
+	struct vdec_vp9_slice_init_vsi *vsi;
+	int ret;
+
+	instance = kzalloc(sizeof(*instance), GFP_KERNEL);
+	if (!instance)
+		return -ENOMEM;
+
+	instance->ctx = ctx;
+	instance->vpu.id = SCP_IPI_VDEC_LAT;
+	instance->vpu.core_id = SCP_IPI_VDEC_CORE;
+	instance->vpu.ctx = ctx;
+	instance->vpu.codec_type = ctx->current_codec;
+
+	ret = vpu_dec_init(&instance->vpu);
+	if (ret) {
+		mtk_vdec_err(ctx, "failed to init vpu dec, ret %d\n", ret);
+		goto error_vpu_init;
+	}
+
+	/* init vsi and global flags */
+
+	vsi = instance->vpu.vsi;
+	if (!vsi) {
+		mtk_vdec_err(ctx, "failed to get VP9 vsi\n");
+		ret = -EINVAL;
+		goto error_vsi;
+	}
+	instance->init_vsi = vsi;
+	instance->core_vsi = mtk_vcodec_fw_map_dm_addr(ctx->dev->fw_handler,
+						       (u32)vsi->core_vsi);
+	if (!instance->core_vsi) {
+		mtk_vdec_err(ctx, "failed to get VP9 core vsi\n");
+		ret = -EINVAL;
+		goto error_vsi;
+	}
+
+	instance->irq = 1;
+
+	ret = vdec_vp9_slice_init_default_frame_ctx(instance);
+	if (ret)
+		goto error_default_frame_ctx;
+
+	ctx->drv_handle = instance;
+
+	return 0;
+
+error_default_frame_ctx:
+error_vsi:
+	vpu_dec_deinit(&instance->vpu);
+error_vpu_init:
+	kfree(instance);
+	return ret;
+}
+
+static void vdec_vp9_slice_deinit(void *h_vdec)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+
+	if (!instance)
+		return;
+
+	vpu_dec_deinit(&instance->vpu);
+	vdec_vp9_slice_free_working_buffer(instance);
+	vdec_msg_queue_deinit(&instance->ctx->msg_queue, instance->ctx);
+	kfree(instance);
+}
+
+static int vdec_vp9_slice_flush(void *h_vdec, struct mtk_vcodec_mem *bs,
+				struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+
+	mtk_vdec_debug(instance->ctx, "flush ...\n");
+	if (instance->ctx->dev->vdec_pdata->hw_arch != MTK_VDEC_PURE_SINGLE_CORE)
+		vdec_msg_queue_wait_lat_buf_full(&instance->ctx->msg_queue);
+	return vpu_dec_reset(&instance->vpu);
+}
+
+static void vdec_vp9_slice_get_pic_info(struct vdec_vp9_slice_instance *instance)
+{
+	struct mtk_vcodec_dec_ctx *ctx = instance->ctx;
+	unsigned int data[3];
+
+	mtk_vdec_debug(instance->ctx, "w %u h %u\n", ctx->picinfo.pic_w, ctx->picinfo.pic_h);
+
+	data[0] = ctx->picinfo.pic_w;
+	data[1] = ctx->picinfo.pic_h;
+	data[2] = ctx->capture_fourcc;
+	vpu_dec_get_param(&instance->vpu, data, 3, GET_PARAM_PIC_INFO);
+
+	ctx->picinfo.buf_w = ALIGN(ctx->picinfo.pic_w, 64);
+	ctx->picinfo.buf_h = ALIGN(ctx->picinfo.pic_h, 64);
+	ctx->picinfo.fb_sz[0] = instance->vpu.fb_sz[0];
+	ctx->picinfo.fb_sz[1] = instance->vpu.fb_sz[1];
+}
+
+static void vdec_vp9_slice_get_dpb_size(struct vdec_vp9_slice_instance *instance,
+					unsigned int *dpb_sz)
+{
+	/* refer VP9 specification */
+	*dpb_sz = 9;
+}
+
+static int vdec_vp9_slice_get_param(void *h_vdec, enum vdec_get_param_type type, void *out)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+
+	switch (type) {
+	case GET_PARAM_PIC_INFO:
+		vdec_vp9_slice_get_pic_info(instance);
+		break;
+	case GET_PARAM_DPB_SIZE:
+		vdec_vp9_slice_get_dpb_size(instance, out);
+		break;
+	case GET_PARAM_CROP_INFO:
+		mtk_vdec_debug(instance->ctx, "No need to get vp9 crop information.");
+		break;
+	default:
+		mtk_vdec_err(instance->ctx, "invalid get parameter type=%d\n", type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int vdec_vp9_slice_single_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+					struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+	struct vdec_vp9_slice_pfc *pfc = &instance->sc_pfc;
+	struct vdec_vp9_slice_vsi *vsi;
+	struct mtk_vcodec_dec_ctx *ctx;
+	int ret;
+
+	if (!instance || !instance->ctx)
+		return -EINVAL;
+	ctx = instance->ctx;
+
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vdec_vp9_slice_flush(h_vdec, bs, fb, res_chg);
+
+	fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx);
+	if (!fb)
+		return -EBUSY;
+
+	vsi = &pfc->vsi;
+
+	ret = vdec_vp9_slice_setup_single(instance, bs, fb, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "Failed to setup VP9 single ret %d\n", ret);
+		return ret;
+	}
+	vdec_vp9_slice_vsi_to_remote(vsi, instance->vsi);
+
+	ret = vpu_dec_start(&instance->vpu, NULL, 0);
+	if (ret) {
+		mtk_vdec_err(ctx, "Failed to dec VP9 ret %d\n", ret);
+		return ret;
+	}
+
+	ret = mtk_vcodec_wait_for_done_ctx(ctx,	MTK_INST_IRQ_RECEIVED,
+					   WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE);
+	/* update remote vsi if decode timeout */
+	if (ret) {
+		mtk_vdec_err(ctx, "VP9 decode timeout %d\n", ret);
+		WRITE_ONCE(instance->vsi->state.timeout, 1);
+	}
+
+	vpu_dec_end(&instance->vpu);
+
+	vdec_vp9_slice_vsi_from_remote(vsi, instance->vsi, 0);
+	ret = vdec_vp9_slice_update_single(instance, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "VP9 decode error: %d\n", ret);
+		return ret;
+	}
+
+	instance->ctx->decoded_frame_cnt++;
+	return 0;
+}
+
+static int vdec_vp9_slice_lat_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				     struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+	struct vdec_lat_buf *lat_buf;
+	struct vdec_vp9_slice_pfc *pfc;
+	struct vdec_vp9_slice_vsi *vsi;
+	struct mtk_vcodec_dec_ctx *ctx;
+	int ret;
+
+	if (!instance || !instance->ctx)
+		return -EINVAL;
+	ctx = instance->ctx;
+
+	/* init msgQ for the first time */
+	if (vdec_msg_queue_init(&ctx->msg_queue, ctx,
+				vdec_vp9_slice_core_decode,
+				sizeof(*pfc)))
+		return -ENOMEM;
+
+	/* bs NULL means flush decoder */
+	if (!bs)
+		return vdec_vp9_slice_flush(h_vdec, bs, fb, res_chg);
+
+	lat_buf = vdec_msg_queue_dqbuf(&instance->ctx->msg_queue.lat_ctx);
+	if (!lat_buf) {
+		mtk_vdec_debug(ctx, "Failed to get VP9 lat buf\n");
+		return -EAGAIN;
+	}
+	pfc = (struct vdec_vp9_slice_pfc *)lat_buf->private_data;
+	if (!pfc) {
+		ret = -EINVAL;
+		goto err_free_fb_out;
+	}
+	vsi = &pfc->vsi;
+
+	ret = vdec_vp9_slice_setup_lat(instance, bs, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "Failed to setup VP9 lat ret %d\n", ret);
+		goto err_free_fb_out;
+	}
+	vdec_vp9_slice_vsi_to_remote(vsi, instance->vsi);
+
+	ret = vpu_dec_start(&instance->vpu, NULL, 0);
+	if (ret) {
+		mtk_vdec_err(ctx, "Failed to dec VP9 ret %d\n", ret);
+		goto err_free_fb_out;
+	}
+
+	if (instance->irq) {
+		ret = mtk_vcodec_wait_for_done_ctx(ctx,	MTK_INST_IRQ_RECEIVED,
+						   WAIT_INTR_TIMEOUT_MS, MTK_VDEC_LAT0);
+		/* update remote vsi if decode timeout */
+		if (ret) {
+			mtk_vdec_err(ctx, "VP9 decode timeout %d pic %d\n", ret, pfc->seq);
+			WRITE_ONCE(instance->vsi->state.timeout, 1);
+		}
+		vpu_dec_end(&instance->vpu);
+	}
+
+	vdec_vp9_slice_vsi_from_remote(vsi, instance->vsi, 0);
+	ret = vdec_vp9_slice_update_lat(instance, lat_buf, pfc);
+
+	/* LAT trans full, no more UBE or decode timeout */
+	if (ret) {
+		mtk_vdec_err(ctx, "VP9 decode error: %d\n", ret);
+		goto err_free_fb_out;
+	}
+
+	mtk_vdec_debug(ctx, "lat dma addr: 0x%lx 0x%lx\n",
+		       (unsigned long)pfc->vsi.trans.dma_addr,
+		       (unsigned long)pfc->vsi.trans.dma_addr_end);
+
+	vdec_msg_queue_update_ube_wptr(&ctx->msg_queue,
+				       vsi->trans.dma_addr_end +
+				       ctx->msg_queue.wdma_addr.dma_addr);
+	vdec_msg_queue_qbuf(&ctx->msg_queue.core_ctx, lat_buf);
+
+	return 0;
+err_free_fb_out:
+	vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf);
+	return ret;
+}
+
+static int vdec_vp9_slice_decode(void *h_vdec, struct mtk_vcodec_mem *bs,
+				 struct vdec_fb *fb, bool *res_chg)
+{
+	struct vdec_vp9_slice_instance *instance = h_vdec;
+	int ret;
+
+	if (instance->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_PURE_SINGLE_CORE)
+		ret = vdec_vp9_slice_single_decode(h_vdec, bs, fb, res_chg);
+	else
+		ret = vdec_vp9_slice_lat_decode(h_vdec, bs, fb, res_chg);
+
+	return ret;
+}
+
+static int vdec_vp9_slice_core_decode(struct vdec_lat_buf *lat_buf)
+{
+	struct vdec_vp9_slice_instance *instance;
+	struct vdec_vp9_slice_pfc *pfc;
+	struct mtk_vcodec_dec_ctx *ctx = NULL;
+	struct vdec_fb *fb = NULL;
+	int ret = -EINVAL;
+
+	if (!lat_buf)
+		goto err;
+
+	pfc = lat_buf->private_data;
+	ctx = lat_buf->ctx;
+	if (!pfc || !ctx)
+		goto err;
+
+	instance = ctx->drv_handle;
+	if (!instance)
+		goto err;
+
+	fb = ctx->dev->vdec_pdata->get_cap_buffer(ctx);
+	if (!fb) {
+		ret = -EBUSY;
+		goto err;
+	}
+
+	ret = vdec_vp9_slice_setup_core(instance, fb, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "vdec_vp9_slice_setup_core\n");
+		goto err;
+	}
+	vdec_vp9_slice_vsi_to_remote(&pfc->vsi, instance->core_vsi);
+
+	ret = vpu_dec_core(&instance->vpu);
+	if (ret) {
+		mtk_vdec_err(ctx, "vpu_dec_core\n");
+		goto err;
+	}
+
+	if (instance->irq) {
+		ret = mtk_vcodec_wait_for_done_ctx(ctx, MTK_INST_IRQ_RECEIVED,
+						   WAIT_INTR_TIMEOUT_MS, MTK_VDEC_CORE);
+		/* update remote vsi if decode timeout */
+		if (ret) {
+			mtk_vdec_err(ctx, "VP9 core timeout pic %d\n", pfc->seq);
+			WRITE_ONCE(instance->core_vsi->state.timeout, 1);
+		}
+		vpu_dec_core_end(&instance->vpu);
+	}
+
+	vdec_vp9_slice_vsi_from_remote(&pfc->vsi, instance->core_vsi, 1);
+	ret = vdec_vp9_slice_update_core(instance, lat_buf, pfc);
+	if (ret) {
+		mtk_vdec_err(ctx, "vdec_vp9_slice_update_core\n");
+		goto err;
+	}
+
+	pfc->vsi.trans.dma_addr_end += ctx->msg_queue.wdma_addr.dma_addr;
+	mtk_vdec_debug(ctx, "core dma_addr_end 0x%lx\n",
+		       (unsigned long)pfc->vsi.trans.dma_addr_end);
+	vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end);
+	ctx->dev->vdec_pdata->cap_to_disp(ctx, 0, lat_buf->src_buf_req);
+
+	return 0;
+
+err:
+	if (ctx && pfc) {
+		/* always update read pointer */
+		vdec_msg_queue_update_ube_rptr(&ctx->msg_queue, pfc->vsi.trans.dma_addr_end);
+
+		if (fb)
+			ctx->dev->vdec_pdata->cap_to_disp(ctx, 1, lat_buf->src_buf_req);
+	}
+	return ret;
+}
+
+const struct vdec_common_if vdec_vp9_slice_lat_if = {
+	.init		= vdec_vp9_slice_init,
+	.decode		= vdec_vp9_slice_decode,
+	.get_param	= vdec_vp9_slice_get_param,
+	.deinit		= vdec_vp9_slice_deinit,
+};
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h
new file mode 100644
index 0000000000..f6abb93652
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_base.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ */
+
+#ifndef _VDEC_DRV_BASE_
+#define _VDEC_DRV_BASE_
+
+#include "vdec_drv_if.h"
+
+struct vdec_common_if {
+	/**
+	 * (*init)() - initialize decode driver
+	 * @ctx     : [in] mtk v4l2 context
+	 * @h_vdec  : [out] driver handle
+	 */
+	int (*init)(struct mtk_vcodec_dec_ctx *ctx);
+
+	/**
+	 * (*decode)() - trigger decode
+	 * @h_vdec  : [in] driver handle
+	 * @bs      : [in] input bitstream
+	 * @fb      : [in] frame buffer to store decoded frame
+	 * @res_chg : [out] resolution change happen
+	 */
+	int (*decode)(void *h_vdec, struct mtk_vcodec_mem *bs,
+		      struct vdec_fb *fb, bool *res_chg);
+
+	/**
+	 * (*get_param)() - get driver's parameter
+	 * @h_vdec : [in] driver handle
+	 * @type   : [in] input parameter type
+	 * @out    : [out] buffer to store query result
+	 */
+	int (*get_param)(void *h_vdec, enum vdec_get_param_type type,
+			 void *out);
+
+	/**
+	 * (*deinit)() - deinitialize driver.
+	 * @h_vdec : [in] driver handle to be deinit
+	 */
+	void (*deinit)(void *h_vdec);
+};
+
+#endif
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c
new file mode 100644
index 0000000000..d0b459b160
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.c
@@ -0,0 +1,128 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ *         Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#include <linux/interrupt.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+
+#include "vdec_drv_if.h"
+#include "mtk_vcodec_dec.h"
+#include "vdec_drv_base.h"
+#include "mtk_vcodec_dec_pm.h"
+
+int vdec_if_init(struct mtk_vcodec_dec_ctx *ctx, unsigned int fourcc)
+{
+	enum mtk_vdec_hw_arch hw_arch = ctx->dev->vdec_pdata->hw_arch;
+	int ret = 0;
+
+	switch (fourcc) {
+	case V4L2_PIX_FMT_H264_SLICE:
+		if (!ctx->dev->vdec_pdata->is_subdev_supported) {
+			ctx->dec_if = &vdec_h264_slice_if;
+			ctx->hw_id = MTK_VDEC_CORE;
+		} else {
+			ctx->dec_if = &vdec_h264_slice_multi_if;
+			ctx->hw_id = IS_VDEC_LAT_ARCH(hw_arch) ? MTK_VDEC_LAT0 : MTK_VDEC_CORE;
+		}
+		break;
+	case V4L2_PIX_FMT_H264:
+		ctx->dec_if = &vdec_h264_if;
+		ctx->hw_id = MTK_VDEC_CORE;
+		break;
+	case V4L2_PIX_FMT_VP8_FRAME:
+		ctx->dec_if = &vdec_vp8_slice_if;
+		ctx->hw_id = MTK_VDEC_CORE;
+		break;
+	case V4L2_PIX_FMT_VP8:
+		ctx->dec_if = &vdec_vp8_if;
+		ctx->hw_id = MTK_VDEC_CORE;
+		break;
+	case V4L2_PIX_FMT_VP9:
+		ctx->dec_if = &vdec_vp9_if;
+		ctx->hw_id = MTK_VDEC_CORE;
+		break;
+	case V4L2_PIX_FMT_VP9_FRAME:
+		ctx->dec_if = &vdec_vp9_slice_lat_if;
+		ctx->hw_id = IS_VDEC_LAT_ARCH(hw_arch) ? MTK_VDEC_LAT0 : MTK_VDEC_CORE;
+		break;
+	case V4L2_PIX_FMT_HEVC_SLICE:
+		ctx->dec_if = &vdec_hevc_slice_multi_if;
+		ctx->hw_id = MTK_VDEC_LAT0;
+		break;
+	case V4L2_PIX_FMT_AV1_FRAME:
+		ctx->dec_if = &vdec_av1_slice_lat_if;
+		ctx->hw_id = MTK_VDEC_LAT0;
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id);
+	ret = ctx->dec_if->init(ctx);
+	mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id);
+
+	return ret;
+}
+
+int vdec_if_decode(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_mem *bs,
+		   struct vdec_fb *fb, bool *res_chg)
+{
+	int ret = 0;
+
+	if (bs) {
+		if ((bs->dma_addr & 63) != 0) {
+			mtk_v4l2_vdec_err(ctx, "bs dma_addr should 64 byte align");
+			return -EINVAL;
+		}
+	}
+
+	if (fb) {
+		if (((fb->base_y.dma_addr & 511) != 0) ||
+		    ((fb->base_c.dma_addr & 511) != 0)) {
+			mtk_v4l2_vdec_err(ctx, "frame buffer dma_addr should 512 byte align");
+			return -EINVAL;
+		}
+	}
+
+	if (!ctx->drv_handle)
+		return -EIO;
+
+	mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id);
+	mtk_vcodec_set_curr_ctx(ctx->dev, ctx, ctx->hw_id);
+	ret = ctx->dec_if->decode(ctx->drv_handle, bs, fb, res_chg);
+	mtk_vcodec_set_curr_ctx(ctx->dev, NULL, ctx->hw_id);
+	mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id);
+
+	return ret;
+}
+
+int vdec_if_get_param(struct mtk_vcodec_dec_ctx *ctx, enum vdec_get_param_type type,
+		      void *out)
+{
+	int ret = 0;
+
+	if (!ctx->drv_handle)
+		return -EIO;
+
+	mtk_vdec_lock(ctx);
+	ret = ctx->dec_if->get_param(ctx->drv_handle, type, out);
+	mtk_vdec_unlock(ctx);
+
+	return ret;
+}
+
+void vdec_if_deinit(struct mtk_vcodec_dec_ctx *ctx)
+{
+	if (!ctx->drv_handle)
+		return;
+
+	mtk_vcodec_dec_enable_hardware(ctx, ctx->hw_id);
+	ctx->dec_if->deinit(ctx->drv_handle);
+	mtk_vcodec_dec_disable_hardware(ctx, ctx->hw_id);
+
+	ctx->drv_handle = NULL;
+}
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h
new file mode 100644
index 0000000000..bfd297c968
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_drv_if.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ *		   Tiffany Lin <tiffany.lin@mediatek.com>
+ */
+
+#ifndef _VDEC_DRV_IF_H_
+#define _VDEC_DRV_IF_H_
+
+#include "mtk_vcodec_dec.h"
+
+
+/**
+ * enum vdec_fb_status  - decoder frame buffer status
+ * @FB_ST_NORMAL: initial state
+ * @FB_ST_DISPLAY: frame buffer is ready to be displayed
+ * @FB_ST_FREE: frame buffer is not used by decoder any more
+ */
+enum vdec_fb_status {
+	FB_ST_NORMAL		= 0,
+	FB_ST_DISPLAY		= (1 << 0),
+	FB_ST_FREE		= (1 << 1)
+};
+
+/* For GET_PARAM_DISP_FRAME_BUFFER and GET_PARAM_FREE_FRAME_BUFFER,
+ * the caller does not own the returned buffer. The buffer will not be
+ *				released before vdec_if_deinit.
+ * GET_PARAM_DISP_FRAME_BUFFER	: get next displayable frame buffer,
+ *				struct vdec_fb**
+ * GET_PARAM_FREE_FRAME_BUFFER	: get non-referenced framebuffer, vdec_fb**
+ * GET_PARAM_PIC_INFO		: get picture info, struct vdec_pic_info*
+ * GET_PARAM_CROP_INFO		: get crop info, struct v4l2_crop*
+ * GET_PARAM_DPB_SIZE		: get dpb size, unsigned int*
+ */
+enum vdec_get_param_type {
+	GET_PARAM_DISP_FRAME_BUFFER,
+	GET_PARAM_FREE_FRAME_BUFFER,
+	GET_PARAM_PIC_INFO,
+	GET_PARAM_CROP_INFO,
+	GET_PARAM_DPB_SIZE
+};
+
+/**
+ * struct vdec_fb_node  - decoder frame buffer node
+ * @list	: list to hold this node
+ * @fb	: point to frame buffer (vdec_fb), fb could point to frame buffer and
+ *	working buffer this is for maintain buffers in different state
+ */
+struct vdec_fb_node {
+	struct list_head list;
+	struct vdec_fb *fb;
+};
+
+extern const struct vdec_common_if vdec_h264_if;
+extern const struct vdec_common_if vdec_h264_slice_if;
+extern const struct vdec_common_if vdec_h264_slice_multi_if;
+extern const struct vdec_common_if vdec_vp8_if;
+extern const struct vdec_common_if vdec_vp8_slice_if;
+extern const struct vdec_common_if vdec_vp9_if;
+extern const struct vdec_common_if vdec_vp9_slice_lat_if;
+extern const struct vdec_common_if vdec_hevc_slice_multi_if;
+extern const struct vdec_common_if vdec_av1_slice_lat_if;
+
+/**
+ * vdec_if_init() - initialize decode driver
+ * @ctx	: [in] v4l2 context
+ * @fourcc	: [in] video format fourcc, V4L2_PIX_FMT_H264/VP8/VP9..
+ */
+int vdec_if_init(struct mtk_vcodec_dec_ctx *ctx, unsigned int fourcc);
+
+/**
+ * vdec_if_deinit() - deinitialize decode driver
+ * @ctx	: [in] v4l2 context
+ *
+ */
+void vdec_if_deinit(struct mtk_vcodec_dec_ctx *ctx);
+
+/**
+ * vdec_if_decode() - trigger decode
+ * @ctx	: [in] v4l2 context
+ * @bs	: [in] input bitstream
+ * @fb	: [in] frame buffer to store decoded frame, when null means parse
+ *	header only
+ * @res_chg	: [out] resolution change happens if current bs have different
+ *	picture width/height
+ * Note: To flush the decoder when reaching EOF, set input bitstream as NULL.
+ *
+ * Return: 0 on success. -EIO on unrecoverable error.
+ */
+int vdec_if_decode(struct mtk_vcodec_dec_ctx *ctx, struct mtk_vcodec_mem *bs,
+		   struct vdec_fb *fb, bool *res_chg);
+
+/**
+ * vdec_if_get_param() - get driver's parameter
+ * @ctx	: [in] v4l2 context
+ * @type	: [in] input parameter type
+ * @out	: [out] buffer to store query result
+ */
+int vdec_if_get_param(struct mtk_vcodec_dec_ctx *ctx, enum vdec_get_param_type type,
+		      void *out);
+
+#endif
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h
new file mode 100644
index 0000000000..47070be2a9
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_ipi_msg.h
@@ -0,0 +1,153 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ */
+
+#ifndef _VDEC_IPI_MSG_H_
+#define _VDEC_IPI_MSG_H_
+
+/*
+ * enum vdec_ipi_msgid - message id between AP and VPU
+ * @AP_IPIMSG_XXX	: AP to VPU cmd message id
+ * @VPU_IPIMSG_XXX_ACK	: VPU ack AP cmd message id
+ */
+enum vdec_ipi_msgid {
+	AP_IPIMSG_DEC_INIT = 0xA000,
+	AP_IPIMSG_DEC_START = 0xA001,
+	AP_IPIMSG_DEC_END = 0xA002,
+	AP_IPIMSG_DEC_DEINIT = 0xA003,
+	AP_IPIMSG_DEC_RESET = 0xA004,
+	AP_IPIMSG_DEC_CORE = 0xA005,
+	AP_IPIMSG_DEC_CORE_END = 0xA006,
+	AP_IPIMSG_DEC_GET_PARAM = 0xA007,
+
+	VPU_IPIMSG_DEC_INIT_ACK = 0xB000,
+	VPU_IPIMSG_DEC_START_ACK = 0xB001,
+	VPU_IPIMSG_DEC_END_ACK = 0xB002,
+	VPU_IPIMSG_DEC_DEINIT_ACK = 0xB003,
+	VPU_IPIMSG_DEC_RESET_ACK = 0xB004,
+	VPU_IPIMSG_DEC_CORE_ACK = 0xB005,
+	VPU_IPIMSG_DEC_CORE_END_ACK = 0xB006,
+	VPU_IPIMSG_DEC_GET_PARAM_ACK = 0xB007,
+};
+
+/**
+ * struct vdec_ap_ipi_cmd - generic AP to VPU ipi command format
+ * @msg_id	: vdec_ipi_msgid
+ * @vpu_inst_addr : VPU decoder instance address. Used if ABI version < 2.
+ * @inst_id     : instance ID. Used if the ABI version >= 2.
+ * @codec_type	: codec fourcc
+ * @reserved	: reserved param
+ */
+struct vdec_ap_ipi_cmd {
+	uint32_t msg_id;
+	union {
+		uint32_t vpu_inst_addr;
+		uint32_t inst_id;
+	};
+	u32 codec_type;
+	u32 reserved;
+};
+
+/**
+ * struct vdec_vpu_ipi_ack - generic VPU to AP ipi command format
+ * @msg_id	: vdec_ipi_msgid
+ * @status	: VPU exeuction result
+ * @ap_inst_addr	: AP video decoder instance address
+ */
+struct vdec_vpu_ipi_ack {
+	uint32_t msg_id;
+	int32_t status;
+	uint64_t ap_inst_addr;
+};
+
+/**
+ * struct vdec_ap_ipi_init - for AP_IPIMSG_DEC_INIT
+ * @msg_id	: AP_IPIMSG_DEC_INIT
+ * @codec_type	: codec fourcc
+ * @ap_inst_addr	: AP video decoder instance address
+ */
+struct vdec_ap_ipi_init {
+	uint32_t msg_id;
+	u32 codec_type;
+	uint64_t ap_inst_addr;
+};
+
+/**
+ * struct vdec_ap_ipi_dec_start - for AP_IPIMSG_DEC_START
+ * @msg_id	: AP_IPIMSG_DEC_START
+ * @vpu_inst_addr : VPU decoder instance address. Used if ABI version < 2.
+ * @inst_id     : instance ID. Used if the ABI version >= 2.
+ * @data	: Header info
+ *	H264 decoder [0]:buf_sz [1]:nal_start
+ *	VP8 decoder  [0]:width/height
+ *	VP9 decoder  [0]:profile, [1][2] width/height
+ * @codec_type	: codec fourcc
+ */
+struct vdec_ap_ipi_dec_start {
+	uint32_t msg_id;
+	union {
+		uint32_t vpu_inst_addr;
+		uint32_t inst_id;
+	};
+	uint32_t data[3];
+	u32 codec_type;
+};
+
+/**
+ * struct vdec_vpu_ipi_init_ack - for VPU_IPIMSG_DEC_INIT_ACK
+ * @msg_id	: VPU_IPIMSG_DEC_INIT_ACK
+ * @status	: VPU exeuction result
+ * @ap_inst_addr	: AP vcodec_vpu_inst instance address
+ * @vpu_inst_addr	: VPU decoder instance address
+ * @vdec_abi_version:	ABI version of the firmware. Kernel can use it to
+ *			ensure that it is compatible with the firmware.
+ *			This field is not valid for MT8173 and must not be
+ *			accessed for this chip.
+ * @inst_id     : instance ID. Valid only if the ABI version >= 2.
+ */
+struct vdec_vpu_ipi_init_ack {
+	uint32_t msg_id;
+	int32_t status;
+	uint64_t ap_inst_addr;
+	uint32_t vpu_inst_addr;
+	uint32_t vdec_abi_version;
+	uint32_t inst_id;
+};
+
+/**
+ * struct vdec_ap_ipi_get_param - for AP_IPIMSG_DEC_GET_PARAM
+ * @msg_id	: AP_IPIMSG_DEC_GET_PARAM
+ * @inst_id     : instance ID. Used if the ABI version >= 2.
+ * @data	: picture information
+ * @param_type	: get param type
+ * @codec_type	: Codec fourcc
+ */
+struct vdec_ap_ipi_get_param {
+	u32 msg_id;
+	u32 inst_id;
+	u32 data[4];
+	u32 param_type;
+	u32 codec_type;
+};
+
+/**
+ * struct vdec_vpu_ipi_get_param_ack - for VPU_IPIMSG_DEC_GET_PARAM_ACK
+ * @msg_id	: VPU_IPIMSG_DEC_GET_PARAM_ACK
+ * @status	: VPU execution result
+ * @ap_inst_addr	: AP vcodec_vpu_inst instance address
+ * @data     : picture information from SCP.
+ * @param_type	: get param type
+ * @reserved : reserved param
+ */
+struct vdec_vpu_ipi_get_param_ack {
+	u32 msg_id;
+	s32 status;
+	u64 ap_inst_addr;
+	u32 data[4];
+	u32 param_type;
+	u32 reserved;
+};
+
+#endif
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c
new file mode 100644
index 0000000000..f283c4703d
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.c
@@ -0,0 +1,372 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#include <linux/freezer.h>
+#include <linux/interrupt.h>
+#include <linux/kthread.h>
+
+#include "mtk_vcodec_dec_drv.h"
+#include "mtk_vcodec_dec_pm.h"
+#include "vdec_msg_queue.h"
+
+#define VDEC_MSG_QUEUE_TIMEOUT_MS 1500
+
+/* the size used to store lat slice header information */
+#define VDEC_LAT_SLICE_HEADER_SZ    (640 * SZ_1K)
+
+/* the size used to store avc error information */
+#define VDEC_ERR_MAP_SZ_AVC         (17 * SZ_1K)
+
+#define VDEC_RD_MV_BUFFER_SZ        (((SZ_4K * 2304 >> 4) + SZ_1K) << 1)
+#define VDEC_LAT_TILE_SZ            (64 * V4L2_AV1_MAX_TILE_COUNT)
+
+/* core will read the trans buffer which decoded by lat to decode again.
+ * The trans buffer size of FHD and 4K bitstreams are different.
+ */
+static int vde_msg_queue_get_trans_size(int width, int height)
+{
+	if (width > 1920 || height > 1088)
+		return 30 * SZ_1M;
+	else
+		return 6 * SZ_1M;
+}
+
+void vdec_msg_queue_init_ctx(struct vdec_msg_queue_ctx *ctx, int hardware_index)
+{
+	init_waitqueue_head(&ctx->ready_to_use);
+	INIT_LIST_HEAD(&ctx->ready_queue);
+	spin_lock_init(&ctx->ready_lock);
+	ctx->ready_num = 0;
+	ctx->hardware_index = hardware_index;
+}
+
+static struct list_head *vdec_get_buf_list(int hardware_index, struct vdec_lat_buf *buf)
+{
+	switch (hardware_index) {
+	case MTK_VDEC_CORE:
+		return &buf->core_list;
+	case MTK_VDEC_LAT0:
+		return &buf->lat_list;
+	default:
+		return NULL;
+	}
+}
+
+static void vdec_msg_queue_inc(struct vdec_msg_queue *msg_queue, int hardware_index)
+{
+	if (hardware_index == MTK_VDEC_CORE)
+		atomic_inc(&msg_queue->core_list_cnt);
+	else
+		atomic_inc(&msg_queue->lat_list_cnt);
+}
+
+static void vdec_msg_queue_dec(struct vdec_msg_queue *msg_queue, int hardware_index)
+{
+	if (hardware_index == MTK_VDEC_CORE)
+		atomic_dec(&msg_queue->core_list_cnt);
+	else
+		atomic_dec(&msg_queue->lat_list_cnt);
+}
+
+int vdec_msg_queue_qbuf(struct vdec_msg_queue_ctx *msg_ctx, struct vdec_lat_buf *buf)
+{
+	struct list_head *head;
+
+	head = vdec_get_buf_list(msg_ctx->hardware_index, buf);
+	if (!head) {
+		mtk_v4l2_vdec_err(buf->ctx, "fail to qbuf: %d", msg_ctx->hardware_index);
+		return -EINVAL;
+	}
+
+	spin_lock(&msg_ctx->ready_lock);
+	list_add_tail(head, &msg_ctx->ready_queue);
+	msg_ctx->ready_num++;
+
+	vdec_msg_queue_inc(&buf->ctx->msg_queue, msg_ctx->hardware_index);
+	if (msg_ctx->hardware_index != MTK_VDEC_CORE) {
+		wake_up_all(&msg_ctx->ready_to_use);
+	} else {
+		if (!(buf->ctx->msg_queue.status & CONTEXT_LIST_QUEUED)) {
+			queue_work(buf->ctx->dev->core_workqueue, &buf->ctx->msg_queue.core_work);
+			buf->ctx->msg_queue.status |= CONTEXT_LIST_QUEUED;
+		}
+	}
+
+	mtk_v4l2_vdec_dbg(3, buf->ctx, "enqueue buf type: %d addr: 0x%p num: %d",
+			  msg_ctx->hardware_index, buf, msg_ctx->ready_num);
+	spin_unlock(&msg_ctx->ready_lock);
+
+	return 0;
+}
+
+static bool vdec_msg_queue_wait_event(struct vdec_msg_queue_ctx *msg_ctx)
+{
+	int ret;
+
+	ret = wait_event_timeout(msg_ctx->ready_to_use,
+				 !list_empty(&msg_ctx->ready_queue),
+				 msecs_to_jiffies(VDEC_MSG_QUEUE_TIMEOUT_MS));
+	if (!ret)
+		return false;
+
+	return true;
+}
+
+struct vdec_lat_buf *vdec_msg_queue_dqbuf(struct vdec_msg_queue_ctx *msg_ctx)
+{
+	struct vdec_lat_buf *buf;
+	struct list_head *head;
+	int ret;
+
+	spin_lock(&msg_ctx->ready_lock);
+	if (list_empty(&msg_ctx->ready_queue)) {
+		spin_unlock(&msg_ctx->ready_lock);
+
+		if (msg_ctx->hardware_index == MTK_VDEC_CORE)
+			return NULL;
+
+		ret = vdec_msg_queue_wait_event(msg_ctx);
+		if (!ret)
+			return NULL;
+		spin_lock(&msg_ctx->ready_lock);
+	}
+
+	if (msg_ctx->hardware_index == MTK_VDEC_CORE)
+		buf = list_first_entry(&msg_ctx->ready_queue,
+				       struct vdec_lat_buf, core_list);
+	else
+		buf = list_first_entry(&msg_ctx->ready_queue,
+				       struct vdec_lat_buf, lat_list);
+
+	head = vdec_get_buf_list(msg_ctx->hardware_index, buf);
+	if (!head) {
+		spin_unlock(&msg_ctx->ready_lock);
+		mtk_v4l2_vdec_err(buf->ctx, "fail to dqbuf: %d", msg_ctx->hardware_index);
+		return NULL;
+	}
+	list_del(head);
+	vdec_msg_queue_dec(&buf->ctx->msg_queue, msg_ctx->hardware_index);
+
+	msg_ctx->ready_num--;
+	mtk_v4l2_vdec_dbg(3, buf->ctx, "dqueue buf type:%d addr: 0x%p num: %d",
+			  msg_ctx->hardware_index, buf, msg_ctx->ready_num);
+	spin_unlock(&msg_ctx->ready_lock);
+
+	return buf;
+}
+
+void vdec_msg_queue_update_ube_rptr(struct vdec_msg_queue *msg_queue, uint64_t ube_rptr)
+{
+	spin_lock(&msg_queue->lat_ctx.ready_lock);
+	msg_queue->wdma_rptr_addr = ube_rptr;
+	mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "update ube rprt (0x%llx)", ube_rptr);
+	spin_unlock(&msg_queue->lat_ctx.ready_lock);
+}
+
+void vdec_msg_queue_update_ube_wptr(struct vdec_msg_queue *msg_queue, uint64_t ube_wptr)
+{
+	spin_lock(&msg_queue->lat_ctx.ready_lock);
+	msg_queue->wdma_wptr_addr = ube_wptr;
+	mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "update ube wprt: (0x%llx 0x%llx) offset: 0x%llx",
+			  msg_queue->wdma_rptr_addr, msg_queue->wdma_wptr_addr,
+			  ube_wptr);
+	spin_unlock(&msg_queue->lat_ctx.ready_lock);
+}
+
+bool vdec_msg_queue_wait_lat_buf_full(struct vdec_msg_queue *msg_queue)
+{
+	if (atomic_read(&msg_queue->lat_list_cnt) == NUM_BUFFER_COUNT) {
+		mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "wait buf full: (%d %d) ready:%d status:%d",
+				  atomic_read(&msg_queue->lat_list_cnt),
+				  atomic_read(&msg_queue->core_list_cnt),
+				  msg_queue->lat_ctx.ready_num, msg_queue->status);
+		return true;
+	}
+
+	msg_queue->flush_done = false;
+	vdec_msg_queue_qbuf(&msg_queue->core_ctx, &msg_queue->empty_lat_buf);
+	wait_event(msg_queue->core_dec_done, msg_queue->flush_done);
+
+	mtk_v4l2_vdec_dbg(3, msg_queue->ctx, "flush done => ready_num:%d status:%d list(%d %d)",
+			  msg_queue->lat_ctx.ready_num, msg_queue->status,
+			  atomic_read(&msg_queue->lat_list_cnt),
+			  atomic_read(&msg_queue->core_list_cnt));
+
+	return false;
+}
+
+void vdec_msg_queue_deinit(struct vdec_msg_queue *msg_queue,
+			   struct mtk_vcodec_dec_ctx *ctx)
+{
+	struct vdec_lat_buf *lat_buf;
+	struct mtk_vcodec_mem *mem;
+	int i;
+
+	mem = &msg_queue->wdma_addr;
+	if (mem->va)
+		mtk_vcodec_mem_free(ctx, mem);
+	for (i = 0; i < NUM_BUFFER_COUNT; i++) {
+		lat_buf = &msg_queue->lat_buf[i];
+
+		mem = &lat_buf->wdma_err_addr;
+		if (mem->va)
+			mtk_vcodec_mem_free(ctx, mem);
+
+		mem = &lat_buf->slice_bc_addr;
+		if (mem->va)
+			mtk_vcodec_mem_free(ctx, mem);
+
+		mem = &lat_buf->rd_mv_addr;
+		if (mem->va)
+			mtk_vcodec_mem_free(ctx, mem);
+
+		mem = &lat_buf->tile_addr;
+		if (mem->va)
+			mtk_vcodec_mem_free(ctx, mem);
+
+		kfree(lat_buf->private_data);
+		lat_buf->private_data = NULL;
+	}
+
+	if (msg_queue->wdma_addr.size)
+		cancel_work_sync(&msg_queue->core_work);
+}
+
+static void vdec_msg_queue_core_work(struct work_struct *work)
+{
+	struct vdec_msg_queue *msg_queue =
+		container_of(work, struct vdec_msg_queue, core_work);
+	struct mtk_vcodec_dec_ctx *ctx =
+		container_of(msg_queue, struct mtk_vcodec_dec_ctx, msg_queue);
+	struct mtk_vcodec_dec_dev *dev = ctx->dev;
+	struct vdec_lat_buf *lat_buf;
+
+	spin_lock(&msg_queue->core_ctx.ready_lock);
+	ctx->msg_queue.status &= ~CONTEXT_LIST_QUEUED;
+	spin_unlock(&msg_queue->core_ctx.ready_lock);
+
+	lat_buf = vdec_msg_queue_dqbuf(&msg_queue->core_ctx);
+	if (!lat_buf)
+		return;
+
+	if (lat_buf->is_last_frame) {
+		ctx->msg_queue.status = CONTEXT_LIST_DEC_DONE;
+		msg_queue->flush_done = true;
+		wake_up(&ctx->msg_queue.core_dec_done);
+
+		return;
+	}
+
+	ctx = lat_buf->ctx;
+	mtk_vcodec_dec_enable_hardware(ctx, MTK_VDEC_CORE);
+	mtk_vcodec_set_curr_ctx(dev, ctx, MTK_VDEC_CORE);
+
+	lat_buf->core_decode(lat_buf);
+
+	mtk_vcodec_set_curr_ctx(dev, NULL, MTK_VDEC_CORE);
+	mtk_vcodec_dec_disable_hardware(ctx, MTK_VDEC_CORE);
+	vdec_msg_queue_qbuf(&ctx->msg_queue.lat_ctx, lat_buf);
+
+	if (!(ctx->msg_queue.status & CONTEXT_LIST_QUEUED) &&
+	    atomic_read(&msg_queue->core_list_cnt)) {
+		spin_lock(&msg_queue->core_ctx.ready_lock);
+		ctx->msg_queue.status |= CONTEXT_LIST_QUEUED;
+		spin_unlock(&msg_queue->core_ctx.ready_lock);
+		queue_work(ctx->dev->core_workqueue, &msg_queue->core_work);
+	}
+}
+
+int vdec_msg_queue_init(struct vdec_msg_queue *msg_queue,
+			struct mtk_vcodec_dec_ctx *ctx, core_decode_cb_t core_decode,
+			int private_size)
+{
+	struct vdec_lat_buf *lat_buf;
+	int i, err;
+
+	/* already init msg queue */
+	if (msg_queue->wdma_addr.size)
+		return 0;
+
+	vdec_msg_queue_init_ctx(&msg_queue->lat_ctx, MTK_VDEC_LAT0);
+	vdec_msg_queue_init_ctx(&msg_queue->core_ctx, MTK_VDEC_CORE);
+	INIT_WORK(&msg_queue->core_work, vdec_msg_queue_core_work);
+
+	atomic_set(&msg_queue->lat_list_cnt, 0);
+	atomic_set(&msg_queue->core_list_cnt, 0);
+	init_waitqueue_head(&msg_queue->core_dec_done);
+	msg_queue->status = CONTEXT_LIST_EMPTY;
+
+	msg_queue->wdma_addr.size =
+		vde_msg_queue_get_trans_size(ctx->picinfo.buf_w,
+					     ctx->picinfo.buf_h);
+	err = mtk_vcodec_mem_alloc(ctx, &msg_queue->wdma_addr);
+	if (err) {
+		mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_addr buf");
+		msg_queue->wdma_addr.size = 0;
+		return -ENOMEM;
+	}
+	msg_queue->wdma_rptr_addr = msg_queue->wdma_addr.dma_addr;
+	msg_queue->wdma_wptr_addr = msg_queue->wdma_addr.dma_addr;
+
+	msg_queue->empty_lat_buf.ctx = ctx;
+	msg_queue->empty_lat_buf.core_decode = NULL;
+	msg_queue->empty_lat_buf.is_last_frame = true;
+
+	msg_queue->ctx = ctx;
+	for (i = 0; i < NUM_BUFFER_COUNT; i++) {
+		lat_buf = &msg_queue->lat_buf[i];
+
+		lat_buf->wdma_err_addr.size = VDEC_ERR_MAP_SZ_AVC;
+		err = mtk_vcodec_mem_alloc(ctx, &lat_buf->wdma_err_addr);
+		if (err) {
+			mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_err_addr buf[%d]", i);
+			goto mem_alloc_err;
+		}
+
+		lat_buf->slice_bc_addr.size = VDEC_LAT_SLICE_HEADER_SZ;
+		err = mtk_vcodec_mem_alloc(ctx, &lat_buf->slice_bc_addr);
+		if (err) {
+			mtk_v4l2_vdec_err(ctx, "failed to allocate wdma_addr buf[%d]", i);
+			goto mem_alloc_err;
+		}
+
+		if (ctx->current_codec == V4L2_PIX_FMT_AV1_FRAME) {
+			lat_buf->rd_mv_addr.size = VDEC_RD_MV_BUFFER_SZ;
+			err = mtk_vcodec_mem_alloc(ctx, &lat_buf->rd_mv_addr);
+			if (err) {
+				mtk_v4l2_vdec_err(ctx, "failed to allocate rd_mv_addr buf[%d]", i);
+				goto mem_alloc_err;
+			}
+
+			lat_buf->tile_addr.size = VDEC_LAT_TILE_SZ;
+			err = mtk_vcodec_mem_alloc(ctx, &lat_buf->tile_addr);
+			if (err) {
+				mtk_v4l2_vdec_err(ctx, "failed to allocate tile_addr buf[%d]", i);
+				goto mem_alloc_err;
+			}
+		}
+
+		lat_buf->private_data = kzalloc(private_size, GFP_KERNEL);
+		if (!lat_buf->private_data) {
+			err = -ENOMEM;
+			goto mem_alloc_err;
+		}
+
+		lat_buf->ctx = ctx;
+		lat_buf->core_decode = core_decode;
+		lat_buf->is_last_frame = false;
+		err = vdec_msg_queue_qbuf(&msg_queue->lat_ctx, lat_buf);
+		if (err) {
+			mtk_v4l2_vdec_err(ctx, "failed to qbuf buf[%d]", i);
+			goto mem_alloc_err;
+		}
+	}
+	return 0;
+
+mem_alloc_err:
+	vdec_msg_queue_deinit(msg_queue, ctx);
+	return err;
+}
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h
new file mode 100644
index 0000000000..1d9beb9e4a
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_msg_queue.h
@@ -0,0 +1,191 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2021 MediaTek Inc.
+ * Author: Yunfei Dong <yunfei.dong@mediatek.com>
+ */
+
+#ifndef _VDEC_MSG_QUEUE_H_
+#define _VDEC_MSG_QUEUE_H_
+
+#include <linux/sched.h>
+#include <linux/semaphore.h>
+#include <linux/slab.h>
+#include <media/videobuf2-v4l2.h>
+
+#define NUM_BUFFER_COUNT 3
+
+struct vdec_lat_buf;
+struct mtk_vcodec_dec_ctx;
+struct mtk_vcodec_dec_dev;
+typedef int (*core_decode_cb_t)(struct vdec_lat_buf *lat_buf);
+
+/**
+ * enum core_ctx_status - Context decode status for core hardwre.
+ * @CONTEXT_LIST_EMPTY: No buffer queued on core hardware(must always be 0)
+ * @CONTEXT_LIST_QUEUED: Buffer queued to core work list
+ * @CONTEXT_LIST_DEC_DONE: context decode done
+ */
+enum core_ctx_status {
+	CONTEXT_LIST_EMPTY = 0,
+	CONTEXT_LIST_QUEUED,
+	CONTEXT_LIST_DEC_DONE,
+};
+
+/**
+ * struct vdec_msg_queue_ctx - represents a queue for buffers ready to be processed
+ * @ready_to_use: ready used queue used to signalize when get a job queue
+ * @ready_queue: list of ready lat buffer queues
+ * @ready_lock: spin lock to protect the lat buffer usage
+ * @ready_num: number of buffers ready to be processed
+ * @hardware_index: hardware id that this queue is used for
+ */
+struct vdec_msg_queue_ctx {
+	wait_queue_head_t ready_to_use;
+	struct list_head ready_queue;
+	/* protect lat buffer */
+	spinlock_t ready_lock;
+	int ready_num;
+	int hardware_index;
+};
+
+/**
+ * struct vdec_lat_buf - lat buffer message used to store lat info for core decode
+ * @wdma_err_addr: wdma error address used for lat hardware
+ * @slice_bc_addr: slice bc address used for lat hardware
+ * @rd_mv_addr:	mv addr for av1 lat hardware output, core hardware input
+ * @tile_addr:	tile buffer for av1 core input
+ * @ts_info: need to set timestamp from output to capture
+ * @src_buf_req: output buffer media request object
+ *
+ * @private_data: shared information used to lat and core hardware
+ * @ctx: mtk vcodec context information
+ * @core_decode: different codec use different decode callback function
+ * @lat_list: add lat buffer to lat head list
+ * @core_list: add lat buffer to core head list
+ *
+ * @is_last_frame: meaning this buffer is the last frame
+ */
+struct vdec_lat_buf {
+	struct mtk_vcodec_mem wdma_err_addr;
+	struct mtk_vcodec_mem slice_bc_addr;
+	struct mtk_vcodec_mem rd_mv_addr;
+	struct mtk_vcodec_mem tile_addr;
+	struct vb2_v4l2_buffer ts_info;
+	struct media_request *src_buf_req;
+
+	void *private_data;
+	struct mtk_vcodec_dec_ctx *ctx;
+	core_decode_cb_t core_decode;
+	struct list_head lat_list;
+	struct list_head core_list;
+
+	bool is_last_frame;
+};
+
+/**
+ * struct vdec_msg_queue - used to store lat buffer message
+ * @lat_buf: lat buffer used to store lat buffer information
+ * @wdma_addr: wdma address used for ube
+ * @wdma_rptr_addr: ube read point
+ * @wdma_wptr_addr: ube write point
+ * @core_work: core hardware work
+ * @lat_ctx: used to store lat buffer list
+ * @core_ctx: used to store core buffer list
+ *
+ * @lat_list_cnt: used to record each instance lat list count
+ * @core_list_cnt: used to record each instance core list count
+ * @flush_done: core flush done status
+ * @empty_lat_buf: the last lat buf used to flush decode
+ * @core_dec_done: core work queue decode done event
+ * @status: current context decode status for core hardware
+ * @ctx: mtk vcodec context information
+ */
+struct vdec_msg_queue {
+	struct vdec_lat_buf lat_buf[NUM_BUFFER_COUNT];
+
+	struct mtk_vcodec_mem wdma_addr;
+	u64 wdma_rptr_addr;
+	u64 wdma_wptr_addr;
+
+	struct work_struct core_work;
+	struct vdec_msg_queue_ctx lat_ctx;
+	struct vdec_msg_queue_ctx core_ctx;
+
+	atomic_t lat_list_cnt;
+	atomic_t core_list_cnt;
+	bool flush_done;
+	struct vdec_lat_buf empty_lat_buf;
+	wait_queue_head_t core_dec_done;
+	int status;
+	struct mtk_vcodec_dec_ctx *ctx;
+};
+
+/**
+ * vdec_msg_queue_init - init lat buffer information.
+ * @msg_queue: used to store the lat buffer information
+ * @ctx: v4l2 ctx
+ * @core_decode: core decode callback for each codec
+ * @private_size: the private data size used to share with core
+ *
+ * Return: returns 0 if init successfully, or fail.
+ */
+int vdec_msg_queue_init(struct vdec_msg_queue *msg_queue,
+			struct mtk_vcodec_dec_ctx *ctx, core_decode_cb_t core_decode,
+			int private_size);
+
+/**
+ * vdec_msg_queue_init_ctx - used to init msg queue context information.
+ * @ctx: message queue context
+ * @hardware_index: hardware index
+ */
+void vdec_msg_queue_init_ctx(struct vdec_msg_queue_ctx *ctx, int hardware_index);
+
+/**
+ * vdec_msg_queue_qbuf - enqueue lat buffer to queue list.
+ * @ctx: message queue context
+ * @buf: current lat buffer
+ *
+ * Return: returns 0 if qbuf successfully, or fail.
+ */
+int vdec_msg_queue_qbuf(struct vdec_msg_queue_ctx *ctx, struct vdec_lat_buf *buf);
+
+/**
+ * vdec_msg_queue_dqbuf - dequeue lat buffer from queue list.
+ * @ctx: message queue context
+ *
+ * Return: returns not null if dq successfully, or fail.
+ */
+struct vdec_lat_buf *vdec_msg_queue_dqbuf(struct vdec_msg_queue_ctx *ctx);
+
+/**
+ * vdec_msg_queue_update_ube_rptr - used to updata the ube read point.
+ * @msg_queue: used to store the lat buffer information
+ * @ube_rptr: current ube read point
+ */
+void vdec_msg_queue_update_ube_rptr(struct vdec_msg_queue *msg_queue, uint64_t ube_rptr);
+
+/**
+ * vdec_msg_queue_update_ube_wptr - used to updata the ube write point.
+ * @msg_queue: used to store the lat buffer information
+ * @ube_wptr: current ube write point
+ */
+void vdec_msg_queue_update_ube_wptr(struct vdec_msg_queue *msg_queue, uint64_t ube_wptr);
+
+/**
+ * vdec_msg_queue_wait_lat_buf_full - used to check whether all lat buffer
+ *                                    in lat list.
+ * @msg_queue: used to store the lat buffer information
+ *
+ * Return: returns true if successfully, or fail.
+ */
+bool vdec_msg_queue_wait_lat_buf_full(struct vdec_msg_queue *msg_queue);
+
+/**
+ * vdec_msg_queue_deinit - deinit lat buffer information.
+ * @msg_queue: used to store the lat buffer information
+ * @ctx: v4l2 ctx
+ */
+void vdec_msg_queue_deinit(struct vdec_msg_queue *msg_queue,
+			   struct mtk_vcodec_dec_ctx *ctx);
+
+#endif
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c
new file mode 100644
index 0000000000..82e57ae983
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c
@@ -0,0 +1,311 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ */
+
+#include "mtk_vcodec_dec_drv.h"
+#include "vdec_drv_if.h"
+#include "vdec_ipi_msg.h"
+#include "vdec_vpu_if.h"
+
+static void handle_init_ack_msg(const struct vdec_vpu_ipi_init_ack *msg)
+{
+	struct vdec_vpu_inst *vpu = (struct vdec_vpu_inst *)
+					(unsigned long)msg->ap_inst_addr;
+
+	mtk_vdec_debug(vpu->ctx, "+ ap_inst_addr = 0x%llx", msg->ap_inst_addr);
+
+	/* mapping VPU address to kernel virtual address */
+	/* the content in vsi is initialized to 0 in VPU */
+	vpu->vsi = mtk_vcodec_fw_map_dm_addr(vpu->ctx->dev->fw_handler,
+					     msg->vpu_inst_addr);
+	vpu->inst_addr = msg->vpu_inst_addr;
+
+	mtk_vdec_debug(vpu->ctx, "- vpu_inst_addr = 0x%x", vpu->inst_addr);
+
+	/* Set default ABI version if dealing with unversioned firmware. */
+	vpu->fw_abi_version = 0;
+	/*
+	 * Instance ID is only used if ABI version >= 2. Initialize it with
+	 * garbage by default.
+	 */
+	vpu->inst_id = 0xdeadbeef;
+
+	/* VPU firmware does not contain a version field. */
+	if (mtk_vcodec_fw_get_type(vpu->ctx->dev->fw_handler) == VPU)
+		return;
+
+	/* Check firmware version. */
+	vpu->fw_abi_version = msg->vdec_abi_version;
+	mtk_vdec_debug(vpu->ctx, "firmware version 0x%x\n", vpu->fw_abi_version);
+	switch (vpu->fw_abi_version) {
+	case 1:
+		break;
+	case 2:
+		vpu->inst_id = msg->inst_id;
+		break;
+	default:
+		mtk_vdec_err(vpu->ctx, "unhandled firmware version 0x%x\n", vpu->fw_abi_version);
+		vpu->failure = 1;
+		break;
+	}
+}
+
+static void handle_get_param_msg_ack(const struct vdec_vpu_ipi_get_param_ack *msg)
+{
+	struct vdec_vpu_inst *vpu = (struct vdec_vpu_inst *)
+					(unsigned long)msg->ap_inst_addr;
+
+	mtk_vdec_debug(vpu->ctx, "+ ap_inst_addr = 0x%llx", msg->ap_inst_addr);
+
+	/* param_type is enum vdec_get_param_type */
+	switch (msg->param_type) {
+	case GET_PARAM_PIC_INFO:
+		vpu->fb_sz[0] = msg->data[0];
+		vpu->fb_sz[1] = msg->data[1];
+		break;
+	default:
+		mtk_vdec_err(vpu->ctx, "invalid get param type=%d", msg->param_type);
+		vpu->failure = 1;
+		break;
+	}
+}
+
+static bool vpu_dec_check_ap_inst(struct mtk_vcodec_dec_dev *dec_dev, struct vdec_vpu_inst *vpu)
+{
+	struct mtk_vcodec_dec_ctx *ctx;
+	int ret = false;
+
+	list_for_each_entry(ctx, &dec_dev->ctx_list, list) {
+		if (!IS_ERR_OR_NULL(ctx) && ctx->vpu_inst == vpu) {
+			ret = true;
+			break;
+		}
+	}
+
+	return ret;
+}
+
+/*
+ * vpu_dec_ipi_handler - Handler for VPU ipi message.
+ *
+ * @data: ipi message
+ * @len : length of ipi message
+ * @priv: callback private data which is passed by decoder when register.
+ *
+ * This function runs in interrupt context and it means there's an IPI MSG
+ * from VPU.
+ */
+static void vpu_dec_ipi_handler(void *data, unsigned int len, void *priv)
+{
+	struct mtk_vcodec_dec_dev *dec_dev;
+	const struct vdec_vpu_ipi_ack *msg = data;
+	struct vdec_vpu_inst *vpu;
+
+	dec_dev = (struct mtk_vcodec_dec_dev *)priv;
+	vpu = (struct vdec_vpu_inst *)(unsigned long)msg->ap_inst_addr;
+	if (!priv || !vpu) {
+		pr_err(MTK_DBG_V4L2_STR "ap_inst_addr is NULL, did the SCP hang or crash?");
+		return;
+	}
+
+	if (!vpu_dec_check_ap_inst(dec_dev, vpu) || msg->msg_id < VPU_IPIMSG_DEC_INIT_ACK ||
+	    msg->msg_id > VPU_IPIMSG_DEC_GET_PARAM_ACK) {
+		mtk_v4l2_vdec_err(vpu->ctx, "vdec msg id not correctly => 0x%x", msg->msg_id);
+		vpu->failure = -EINVAL;
+		goto error;
+	}
+
+	vpu->failure = msg->status;
+	if (msg->status != 0)
+		goto error;
+
+	switch (msg->msg_id) {
+	case VPU_IPIMSG_DEC_INIT_ACK:
+		handle_init_ack_msg(data);
+		break;
+
+	case VPU_IPIMSG_DEC_START_ACK:
+	case VPU_IPIMSG_DEC_END_ACK:
+	case VPU_IPIMSG_DEC_DEINIT_ACK:
+	case VPU_IPIMSG_DEC_RESET_ACK:
+	case VPU_IPIMSG_DEC_CORE_ACK:
+	case VPU_IPIMSG_DEC_CORE_END_ACK:
+		break;
+
+	case VPU_IPIMSG_DEC_GET_PARAM_ACK:
+		handle_get_param_msg_ack(data);
+		break;
+	default:
+		mtk_vdec_err(vpu->ctx, "invalid msg=%X", msg->msg_id);
+		break;
+	}
+
+error:
+	vpu->signaled = 1;
+}
+
+static int vcodec_vpu_send_msg(struct vdec_vpu_inst *vpu, void *msg, int len)
+{
+	int err, id, msgid;
+
+	msgid = *(uint32_t *)msg;
+	mtk_vdec_debug(vpu->ctx, "id=%X", msgid);
+
+	vpu->failure = 0;
+	vpu->signaled = 0;
+
+	if (vpu->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_LAT_SINGLE_CORE) {
+		if (msgid == AP_IPIMSG_DEC_CORE ||
+		    msgid == AP_IPIMSG_DEC_CORE_END)
+			id = vpu->core_id;
+		else
+			id = vpu->id;
+	} else {
+		id = vpu->id;
+	}
+
+	err = mtk_vcodec_fw_ipi_send(vpu->ctx->dev->fw_handler, id, msg,
+				     len, 2000);
+	if (err) {
+		mtk_vdec_err(vpu->ctx, "send fail vpu_id=%d msg_id=%X status=%d",
+			     id, msgid, err);
+		return err;
+	}
+
+	return vpu->failure;
+}
+
+static int vcodec_send_ap_ipi(struct vdec_vpu_inst *vpu, unsigned int msg_id)
+{
+	struct vdec_ap_ipi_cmd msg;
+	int err = 0;
+
+	mtk_vdec_debug(vpu->ctx, "+ id=%X", msg_id);
+
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_id = msg_id;
+	if (vpu->fw_abi_version < 2)
+		msg.vpu_inst_addr = vpu->inst_addr;
+	else
+		msg.inst_id = vpu->inst_id;
+	msg.codec_type = vpu->codec_type;
+
+	err = vcodec_vpu_send_msg(vpu, &msg, sizeof(msg));
+	mtk_vdec_debug(vpu->ctx, "- id=%X ret=%d", msg_id, err);
+	return err;
+}
+
+int vpu_dec_init(struct vdec_vpu_inst *vpu)
+{
+	struct vdec_ap_ipi_init msg;
+	int err;
+
+	init_waitqueue_head(&vpu->wq);
+	vpu->handler = vpu_dec_ipi_handler;
+	vpu->ctx->vpu_inst = vpu;
+
+	err = mtk_vcodec_fw_ipi_register(vpu->ctx->dev->fw_handler, vpu->id,
+					 vpu->handler, "vdec", vpu->ctx->dev);
+	if (err) {
+		mtk_vdec_err(vpu->ctx, "vpu_ipi_register fail status=%d", err);
+		return err;
+	}
+
+	if (vpu->ctx->dev->vdec_pdata->hw_arch == MTK_VDEC_LAT_SINGLE_CORE) {
+		err = mtk_vcodec_fw_ipi_register(vpu->ctx->dev->fw_handler,
+						 vpu->core_id, vpu->handler,
+						 "vdec", vpu->ctx->dev);
+		if (err) {
+			mtk_vdec_err(vpu->ctx, "vpu_ipi_register core fail status=%d", err);
+			return err;
+		}
+	}
+
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_id = AP_IPIMSG_DEC_INIT;
+	msg.ap_inst_addr = (unsigned long)vpu;
+	msg.codec_type = vpu->codec_type;
+
+	mtk_vdec_debug(vpu->ctx, "vdec_inst=%p", vpu);
+
+	err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg));
+	mtk_vdec_debug(vpu->ctx, "- ret=%d", err);
+	return err;
+}
+
+int vpu_dec_start(struct vdec_vpu_inst *vpu, uint32_t *data, unsigned int len)
+{
+	struct vdec_ap_ipi_dec_start msg;
+	int i;
+	int err = 0;
+
+	if (len > ARRAY_SIZE(msg.data)) {
+		mtk_vdec_err(vpu->ctx, "invalid len = %d\n", len);
+		return -EINVAL;
+	}
+
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_id = AP_IPIMSG_DEC_START;
+	if (vpu->fw_abi_version < 2)
+		msg.vpu_inst_addr = vpu->inst_addr;
+	else
+		msg.inst_id = vpu->inst_id;
+
+	for (i = 0; i < len; i++)
+		msg.data[i] = data[i];
+	msg.codec_type = vpu->codec_type;
+
+	err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg));
+	mtk_vdec_debug(vpu->ctx, "- ret=%d", err);
+	return err;
+}
+
+int vpu_dec_get_param(struct vdec_vpu_inst *vpu, uint32_t *data,
+		      unsigned int len, unsigned int param_type)
+{
+	struct vdec_ap_ipi_get_param msg;
+	int err;
+
+	if (len > ARRAY_SIZE(msg.data)) {
+		mtk_vdec_err(vpu->ctx, "invalid len = %d\n", len);
+		return -EINVAL;
+	}
+
+	memset(&msg, 0, sizeof(msg));
+	msg.msg_id = AP_IPIMSG_DEC_GET_PARAM;
+	msg.inst_id = vpu->inst_id;
+	memcpy(msg.data, data, sizeof(unsigned int) * len);
+	msg.param_type = param_type;
+	msg.codec_type = vpu->codec_type;
+
+	err = vcodec_vpu_send_msg(vpu, (void *)&msg, sizeof(msg));
+	mtk_vdec_debug(vpu->ctx, "- ret=%d", err);
+	return err;
+}
+
+int vpu_dec_core(struct vdec_vpu_inst *vpu)
+{
+	return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_CORE);
+}
+
+int vpu_dec_end(struct vdec_vpu_inst *vpu)
+{
+	return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_END);
+}
+
+int vpu_dec_core_end(struct vdec_vpu_inst *vpu)
+{
+	return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_CORE_END);
+}
+
+int vpu_dec_deinit(struct vdec_vpu_inst *vpu)
+{
+	return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_DEINIT);
+}
+
+int vpu_dec_reset(struct vdec_vpu_inst *vpu)
+{
+	return vcodec_send_ap_ipi(vpu, AP_IPIMSG_DEC_RESET);
+}
diff --git a/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h
new file mode 100644
index 0000000000..fbb3f34a73
--- /dev/null
+++ b/drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.h
@@ -0,0 +1,120 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (c) 2016 MediaTek Inc.
+ * Author: PC Chen <pc.chen@mediatek.com>
+ */
+
+#ifndef _VDEC_VPU_IF_H_
+#define _VDEC_VPU_IF_H_
+
+struct mtk_vcodec_dec_ctx;
+
+/**
+ * struct vdec_vpu_inst - VPU instance for video codec
+ * @id          : ipi msg id for each decoder
+ * @core_id     : core id used to separate different hardware
+ * @vsi         : driver structure allocated by VPU side and shared to AP side
+ *                for control and info share
+ * @failure     : VPU execution result status, 0: success, others: fail
+ * @inst_addr	: VPU decoder instance address
+ * @fw_abi_version : ABI version of the firmware.
+ * @inst_id	: if fw_abi_version >= 2, contains the instance ID to be given
+ *                in place of inst_addr in messages.
+ * @signaled    : 1 - Host has received ack message from VPU, 0 - not received
+ * @ctx         : context for v4l2 layer integration
+ * @dev		: platform device of VPU
+ * @wq          : wait queue to wait VPU message ack
+ * @handler     : ipi handler for each decoder
+ * @codec_type     : use codec type to separate different codecs
+ * @capture_type:	used capture type to separate different capture format
+ * @fb_sz  : frame buffer size of each plane
+ */
+struct vdec_vpu_inst {
+	int id;
+	int core_id;
+	void *vsi;
+	int32_t failure;
+	uint32_t inst_addr;
+	uint32_t fw_abi_version;
+	uint32_t inst_id;
+	unsigned int signaled;
+	struct mtk_vcodec_dec_ctx *ctx;
+	wait_queue_head_t wq;
+	mtk_vcodec_ipi_handler handler;
+	unsigned int codec_type;
+	unsigned int capture_type;
+	unsigned int fb_sz[2];
+};
+
+/**
+ * vpu_dec_init - init decoder instance and allocate required resource in VPU.
+ *
+ * @vpu: instance for vdec_vpu_inst
+ */
+int vpu_dec_init(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_start - start decoding, basically the function will be invoked once
+ *                 every frame.
+ *
+ * @vpu : instance for vdec_vpu_inst
+ * @data: meta data to pass bitstream info to VPU decoder
+ * @len : meta data length
+ */
+int vpu_dec_start(struct vdec_vpu_inst *vpu, uint32_t *data, unsigned int len);
+
+/**
+ * vpu_dec_end - end decoding, basically the function will be invoked once
+ *               when HW decoding done interrupt received successfully. The
+ *               decoder in VPU will continue to do reference frame management
+ *               and check if there is a new decoded frame available to display.
+ *
+ * @vpu : instance for vdec_vpu_inst
+ */
+int vpu_dec_end(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_deinit - deinit decoder instance and resource freed in VPU.
+ *
+ * @vpu: instance for vdec_vpu_inst
+ */
+int vpu_dec_deinit(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_reset - reset decoder, use for flush decoder when end of stream or
+ *                 seek. Remainig non displayed frame will be pushed to display.
+ *
+ * @vpu: instance for vdec_vpu_inst
+ */
+int vpu_dec_reset(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_core - core start decoding, basically the function will be invoked once
+ *                 every frame.
+ *
+ * @vpu : instance for vdec_vpu_inst
+ */
+int vpu_dec_core(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_core_end - core end decoding, basically the function will be invoked once
+ *               when core HW decoding done and receive interrupt successfully. The
+ *               decoder in VPU will updata hardware information and deinit hardware
+ *               and check if there is a new decoded frame available to display.
+ *
+ * @vpu : instance for vdec_vpu_inst
+ */
+int vpu_dec_core_end(struct vdec_vpu_inst *vpu);
+
+/**
+ * vpu_dec_get_param - get param from scp
+ *
+ * @vpu : instance for vdec_vpu_inst
+ * @data: meta data to pass bitstream info to VPU decoder
+ * @len : meta data length
+ * @param_type : get param type
+ */
+int vpu_dec_get_param(struct vdec_vpu_inst *vpu, uint32_t *data,
+		      unsigned int len, unsigned int param_type);
+
+#endif